Systems, devices, and methods for tracking and compensating for patient motion during a medical imaging scan
09734589 · 2017-08-15
Assignee
Inventors
- Jeffrey N. Yu (Honolulu, HI)
- William Herman Alameida, Jr. (Honolulu, HI)
- John Arthur Lovberg (San Diego, CA)
- Xiaojiang Jason Pan (San Diego, CA)
- Michael Engelmann (Pukalani, HI)
Cpc classification
A61N5/1049
HUMAN NECESSITIES
A61B5/055
HUMAN NECESSITIES
A61B34/20
HUMAN NECESSITIES
A61B5/0077
HUMAN NECESSITIES
A61B5/721
HUMAN NECESSITIES
International classification
A61N5/10
HUMAN NECESSITIES
A61B5/11
HUMAN NECESSITIES
A61B5/00
HUMAN NECESSITIES
A61B5/055
HUMAN NECESSITIES
G06T7/246
PHYSICS
Abstract
A motion tracking system for dynamic tracking of and compensation for motion of a patient during a magnetic resonance scan comprises a first camera positioned to view an optical marker along a first line of sight; a second camera positioned to view the optical marker along a second line of sight; and a computer system configured to analyze images generated by the first and second cameras to determine changes in position of the optical marker, and to generate tracking data for use by a magnetic resonance scanner to dynamically adjust scans to compensate for the changes in position of the optical marker, wherein the computer system is configured to dynamically adapt its image analysis to utilize images from all cameras that are currently viewing the optical marker.
Claims
1. A motion tracking system for tracking and compensating for motion of a patient during a medical imaging scan, the motion tracking system comprising: an optical marker configured to be attached to the patient, wherein the optical marker comprises a plurality of optically visible landmarks, the plurality of optically visible landmarks defining a plurality of reference points; a plurality of optical detectors, wherein each of the plurality of optical detectors is positioned to view the optical marker along a different line of sight; one or more computer readable storage devices configured to store a plurality of computer executable instructions; and one or more hardware computer processors in communication with the one or more computer readable storage devices and configured to execute the plurality of computer executable instructions in order cause the motion tracking system to: receive a digital image from each of the plurality of optical detectors, wherein each digital image represents a view along a different line of sight; determine, for each digital image, whether the digital image includes a view of the optical marker; determine, for each digital image that includes a view of the optical marker, positions of the plurality of reference points in the digital image; calculate a plurality of baseline attributes related to the plurality of reference points, the plurality of baseline attributes calculated based on the determined positions of the plurality of reference points; and estimate iteratively a three-dimensional pose of the patient, until a measure of error is within a threshold amount, the measure of error calculated based on the plurality of baseline attributes as compared to a plurality of comparison attributes, the plurality of comparison attributes calculated by assuming the object is in an estimated pose.
2. The motion tracking system of claim 1, wherein the motion tracking system is further caused to repeatedly estimate the three-dimensional pose of the patient at a rate of at least 100 Hz.
3. The motion tracking system of claim 1, wherein the plurality of baseline attributes comprises a set of principal quantities, the set of principal quantities comprising values associated with the relationships between the positions of reference points.
4. The motion tracking system of claim 1, wherein determining whether each digital image includes a view of the optical marker comprises: classifying each pixel in the digital image as one of two colors; grouping one or more connected regions, wherein the connected regions are comprised of adjacent pixels with the some color; filtering out one or more connected regions based on a size threshold; computing the centroid of one or more connected regions; and grouping together one or more connected regions by their computed centroids.
5. The motion tracking system of claim 1, wherein the motion tracking system is further caused to remap a coordinate system associated with the estimated three-dimensional pose of the patient.
6. The motion tracking system of claim 4, wherein grouping one or more connected regions comprises: identifying connected regions with centroids occurring within one pixel of each other; and calculating a centroid for each group of the one or more connected regions.
7. The motion tracking system of claim 4, wherein the motion tracking system is further caused to binarize each digital image before classifying each pixel in the digital image as one of two colors.
8. The motion tracking system of claim 6, wherein calculating a centroid for each group of the one or more connected regions comprises averaging the centroid for each connected region in the group of one or more connected regions.
9. The motion tracking system of claim 1, wherein the motion tracking system is further caused to generate tracking data based on the estimated three-dimensional pose of the patient and to transmit the tracking data to a medical imaging scanner controller to enable a medical imaging scanner to dynamically adjust the medical imaging scan to compensate for patient motion.
10. The motion tracking system of claim 9, further comprising: the medical imaging scanner, wherein the medical imaging scanner is configured to dynamically adjust the medical imaging scan to compensate for patient motion based on the tracking data.
11. A motion tracking system for dynamic tracking of and compensation for motion of a patient during a magnetic resonance scan, the motion tracking system comprising: an optical marker configured to be attached to a patient being scanned, wherein the optical marker comprises an optically visible pattern comprising a plurality of reference point locators, each reference point locator configured to define a single reference point, each reference point locator comprising alternating dark and light elliptical shapes centered on the single reference point, wherein at least one of the plurality of reference point locators is larger than at least another of the plurality of reference point locators, and wherein the optically visible pattern is rotationally asymmetrical about an axis normal to a plane of the optically visible pattern; a first camera positioned to view the optical marker along a first line of sight; a second camera positioned to view the optical marker along a second line of sight; and a computer system configured to analyze images generated by the first and second cameras to determine changes in position of the optical marker, and to generate tracking data for use by a magnetic resonance scanner to dynamically adjust scans to compensate for the changes in position of the optical marker, wherein the computer system is configured to dynamically determine whether the first camera, the second camera, or both cameras are currently viewing the optical marker, and wherein the computer system is configured to dynamically adapt its image analysis to utilize images from all cameras that are currently viewing the optical marker.
12. The motion tracking system of claim 11, further comprising: one or more additional cameras each positioned to view the optical marker along a different line of sight, wherein the computer system is further configured to dynamically determine which of all of the cameras are currently viewing the optical marker.
13. The motion tracking system of claim 11, further comprising: a third camera positioned to view the optical marker along a third line of sight; and a fourth camera positioned to view the optical marker along a fourth line of sight, wherein the computer system is further configured to dynamically determine which of all of the cameras are currently viewing the optical marker.
14. The motion tracking system of claim 11, further comprising: one or more additional optical markers configured to be attached to the patient being scanned, wherein the computer system is configured to analyze images generated by all cameras that are currently viewing at least one optical marker to determine changes in position of the optical markers, and to generate tracking data for use by the magnetic resonance scanner to dynamically adjust scans to compensate for the changes in position of the optical markers.
15. The motion tracking system of claim 11, further comprising: the magnetic resonance scanner, wherein the magnetic resonance scanner is configured to dynamically adjust scans to compensate for the changes in position of the optical marker based on the tracking data.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) The foregoing and other features, aspects, and advantages of the present inventions are described in detail below with reference to the drawings of various embodiments, which are intended to illustrate and not to limit the inventions. The drawings comprise the following figures in which:
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
(49)
(50)
(51)
(52)
(53)
(54)
(55)
(56)
(57)
(58)
(59)
(60)
(61)
(62)
(63)
(64)
(65)
(66)
(67)
(68)
(69)
(70)
(71)
(72)
(73)
DETAILED DESCRIPTION
(74) Although several embodiments, examples, and illustrations are disclosed below, it will be understood by those of ordinary skill in the art that the inventions described herein extend beyond the specifically disclosed embodiments, examples, and illustrations and includes other uses of the inventions and obvious modifications and equivalents thereof. Embodiments of the inventions are described with reference to the accompanying figures, wherein like numerals refer to like elements throughout. The terminology used in the description presented herein is not intended to be interpreted in any limited or restrictive manner simply because it is being used in conjunction with a detailed description of certain specific embodiments of the inventions. In addition, embodiments of the inventions can comprise several novel features and no single feature is solely responsible for its desirable attributes or is essential to practicing the inventions herein described.
(75) With the use of diagnostic technologies and therapeutic technologies, it can be advantageous to track for patient movement with a high degree of accuracy. Such high accuracy tracking can improve the imaging quality obtained and produced by diagnostic equipment, such as imaging technologies. Further, the use of high accuracy patient movement tracking technology can improve the application of patient therapies, such as radiation treatment, proton treatment, and the like. By accounting for patient movement with a high degree of accuracy, therapeutic technologies can apply therapies only to the targeted tissue and avoid healthy surrounding tissue.
(76) The embodiments disclosed herein relate to a patient motion tracking system that can track patient movement with translation accuracies of about 0.1 mm and angle accuracies of about 0.1 degrees. As disclosed herein, the system can be configured to utilize a non-stereo approach to determining the 6 degrees of freedom movement of the patient. In an embodiment, the system can comprise two cameras that are positioned orthogonal and perpendicular on a single plane. In an embodiment, the two cameras need not be in a single plane, but are positioned such that the two cameras are not viewing the target from generally the same direction. The system can be configured to compare the appearance of the target on one camera with the other camera while not accounting for which pixel number they fall on in either camera. By comparing the appearance of the target between the two cameras, the system can be configured to extract the 6 degrees of freedom movement on a very small target.
(77) In an embodiment, the system can be configured to extract the movement data based on analyzing the images of the target from the two cameras in order to generate a predicted value for at least one of the variables in the 6 degrees of freedom. For example, the system can be configured to analyze the image of the target and predict a value for the pitch. The system can then be configured to compare the predicted value to the value of the particular variable with that which is shown in the actual image of the target. The system can be configured to repeat this process using an iterative approach to continuously improve the predicted value of one of the variables in the 6 degrees of freedom. The system can be configured to perform this iterative process for each variable in the 6 degrees of freedom.
(78) The foregoing methodology for tracking patient movement can be applied in the diagnostic context as well as in the therapeutic context. For example, as disclosed herein, the system can be configured to track patient movement in order to feed such movement data to an MRI scanner such that the MRI scanner can adjust the focus and position of the scanner in order to produce a clear MRI image of the patient. Further, the system can be configured to connect to therapeutic technologies. For example, the system can be configured to track patient movement in order to direct a therapeutic radiation beam at a diseased tissue region while avoiding surrounding healthy tissue.
(79) There are various technologies for therapeutic radiation and other therapeutics. For example, it can be advantageous in radiation therapy, proton therapy, or other therapies to dynamically apply the radiation to a targeted area in order to account for patient movement. Patient movement can include respiration, twitches or any other voluntary or involuntary movements of the patient. By dynamically and automatically tracking patient movement, radiation therapy, proton therapy, and any other kind of therapy can be applied in a more targeted way, thereby allowing surrounding healthy tissue to be avoided and/or unharmed. The systems disclosed herein can be adapted and configured to track patient translations with accuracies of about 0.1 mm and angle accuracies of about 0.1 degrees in order to better apply radiation therapy, proton therapy, or any other therapy to the targeted tissue or area of the body.
(80) In an embodiment, a system can be configured to utilize optical tracking based on the methods disclosed herein in order to track patient movement and/or or another device, for example, electronics packages that are configured to identify fiducial markers implanted inside a patient. In an embodiment, the system can be configured to utilize the electronics package in order to identify the location of the fiducial markers within the patient. By identifying the location of the fiducial markers, the system needs to identify the location of the electronics package in order to determine the location of the fiducial markers with respect to a scanner and/or a therapeutic equipment device.
(81) The patient tracking movement system, disclosed herein, can be utilized to track periodic involuntary movement of the patient, such as breathing. By tracking the periodic patient movement with a high degree of accuracy, the system can be configured to apply a radiation therapy, a proton therapy, or the like during strategic moments when the target tissue is in a certain position while the patient's involuntary movements continue. Additionally, the system can be configured to track not only normal breathing movement of the patient, but also the system can be configured to track irregular movement of the patient caused by patient activity or based on diseased tissue of the patient. For example, when a patient is running, the ribs of the patient have a larger egression that the system can track in order to continuously identify a target tissue area. In another example, the patient may be suffering from COPD or other breathing disorder or diagrammatic issues. For example, the patient could be suffering from theurofusion, which is water outside the lung that prevents the patient from breathing or a tumor is irritating a lung region thereby preventing normal breathing. The system can be configured to track such irregular patient movements due to such conditions.
(82) In order to apply a therapy, such as radiation therapy, the radiation beam generator must determine the location of the electronics package relative to the beam generator in order to properly direct the radiation therapy to the targeted tissue. Accordingly, it is necessary to track the position of the electronics package relative to the radiation beam generator or other therapeutic equipment. It can be advantageous to track the position of the electronics package with a high degree of accuracy in order to better target the desired tissue. In systems where the electronics package is configured to track the location of fiducial markers implanted within the patient, such systems have two possible sources of error. One source of error can be derived from tracking the position of the fiducial markers using the electronics package and the second source of error can be derived from tracking the position of the electronics package relative to the therapeutic equipment generator. Accordingly, it can be advantageous to identify the position of the electronics package with a high degree of accuracy in order to avoid compounding the sources of error.
(83) Motion Compensation Systems
(84)
(85) In this embodiment, the optical marker 110 is configured to be viewable by each of the two detectors 108. The detectors 108 can be, for example, digital cameras capable of acquiring images of the optical marker 110 and transmitting those images to the motion tracking system 102. In this embodiment, each of the detectors 108 is configured to view the optical marker 110 from along a different line of sight. This can be helpful, for example, to enable the motion tracking system 102 to analyze two dimensional images of the optical marker 110 from different vantage points to help in locating the optical marker 110 to estimate patient motion or pose. In this embodiment, the detectors 108 each are configured to view the optical marker 110 along a line of sight 120 separated from each other by an angle 122. In this embodiment, the angle 122 is approximately 90 degrees. Other angles may be used, such as 30 degrees, 45 degrees, 60 degrees, 70 degrees, etc. In some embodiments, 90 degrees is an optimal angle to enable maximum differentiation of in plane and out of plane motion of the optical marker 110, as further described below with reference to
(86) In some embodiments, the angle 122 may be referred to as a scissor angle. In the embodiment illustrated in
(87) Mirrors or other devices used to redirect a line of sight have both advantages and disadvantages. For example, disadvantages of mirrors include that they could potentially vibrate, potentially introducing error into the object orientation determination process. As another example, the further away a mirror is from a detector, generally the larger the mirror needs to be to enable an equivalent range of vision. Accordingly, it can be advantageous to position a mirror relatively close to a detector to enable the mirror to be relatively small. One advantage of using mirrors or other sight line redirection methods is that a virtual scissor angle can be configured to be closer to an optimal scissor angle of 90°, even when a particular medical imaging scanner configuration may not allow for detectors that are positioned to directly view a marker using a 90° scissor angle. Further, some mirrors are not conductive, which can be advantageous in magnetic resonance imaging, because nonconductive mirrors will not introduce artifacts into MRI images. A digital camera, on the other hand, may include conductive components and/or a wire leading to the detector may include conductive components. When a digital camera and/or its wire are within the medical imaging envelope, they may introduce artifacts into MRI images.
(88) The embodiment of a motion compensation system 100 illustrated in
(89) In the embodiment of a motion compensation system 100 illustrated in
(90)
(91) The optical marker 110 comprises a pattern 202 that defines a reference shape. In this embodiment, the pattern 202 defines a reference shape of an equilateral triangle having sides of approximately 0.5 inches. At each vertex of the equilateral triangle reference shape is a reference point locater 204. In this embodiment, each reference point locator 204 comprises a series of alternating black and white (or dark and light) elliptical shapes, with a centroid of the reference point locater 204 being positioned at the vertex of the equilateral triangle reference shape. In various embodiments, different reference shapes can be used and reference point locators can take various forms, as long as the reference point locators are able to be detected and analyzed by a motion tracking system to determine important points, inflection points, critical points, or vertex points of a reference shape.
(92) In this embodiment, the reference point locators 204 are elliptical in shape and positioned such that they are configured to be visible as a circular pattern from a 45 degree viewing angle. This can be advantageous, because, when used in a system such as the example illustrated in
(93)
(94)
(95) The detectors 108 can comprise, for example, digital cameras. Although in this embodiment there are two detectors 108, various other embodiments may utilize more or fewer detectors based on the application. For example, an embodiment of a motion compensation system may comprise more detectors to increase accuracy of motion tracking and/or to add redundancy to a motion compensation system. For example, a motion compensation system may comprise four detectors, with motion tracking being performed only using two of the detectors at any one time. This may be advantageous, for example, when obstructions may hide an optical marker from view of one or more detectors depending on the position of the object being tracked.
(96) Although in this embodiment and various other embodiments described herein the detectors are optical digital cameras, various other motion compensation systems may utilize detectors other than optical cameras. For example, a detector may be an infrared camera configured to view a target or marker viewable by an infrared camera. In other embodiments, the detectors may comprise laser detectors, sonar detectors, radar detectors, and various other types of detectors capable of locating a marker and/or creating a two dimensional digital image or representation of a marker.
(97) The motion tracking system 102 comprises a tracking engine 304, a calibration engine 306, a controller interface 308, and a motion database 310. The motion database 310 can be configured to store motion tracking information, such as object pose estimates created by the tracking engine 304. In some embodiments, the motion database 310 can be configured to be persistent storage to store the motion information for later retrieval and usage after the completion of an imaging scan. In some embodiments, the motion database 310 comprises a short term memory or buffer or cache to store object pose estimates just temporarily until they are included in motion data and sent to the scanner controller 106 by the controller interface 308.
(98) The calibration engine 306 can be configured to calibrate the motion compensation system. The calibration engine 306 comprises a calibration database 316, a detector calibration filter 318, and a target calibration filter 320. In some embodiments, the calibration engine 306 can be configured to calibrate the system at initial startup or initial system assembly, with calibration not being needed for later operation of the motion compensation system 300. In other embodiments, the calibration engine 306 is utilized for at least some calibration procedures during some or all motion tracking procedures. The detector calibration filter 318 can be configured to calibrate the detectors 108 to the motion compensation system. In some embodiments, the detector calibration filter 318 can be configured to calibrate the detectors by enabling a manufacturer or assembler of the motion compensation system to input information specific to each detector 108, such as focal length, resolution, etc. In some embodiments, the detector calibration filter 318 can be configured to automatically determine some or all of the parameters needed to calibrate a detector 108. The parameters determined in calibration of the detectors can be stored in the calibration database 316 for later use by the tracking engine 304.
(99) The target calibration filter 320 can be configured to calibrate the system to one or more specific targets or markers. For example, the target calibration filter 320 can be configured to, upon initial startup of a motion tracking routine, analyze images received from the detectors 108 to determine a region of interest in the images where it is most likely that the optical marker exists. This information can be stored in the calibration database 316. This initial calibration upon startup of a motion tracking routine can, for example, help to speed up a tracking routine.
(100) The tracking engine 304 comprises a marker location filter 312 and an objection orientation filter 314. The tracking engine 304 can be configured to track the location of one or more optical markers during an imaging scan and to determine an estimate of the pose of the object or patient being scanned. In some embodiments, the marker location filter 312 can be configured to analyze images from the detectors 108 to determine locations of reference points of an optical marker in the 2D images from the detectors. The marker location filter 312 can be configured to analyze these images to determine both the locations of these points and the orientation of the reference shape or shapes formed by the reference points. The object orientation filter 314 can be configured to utilize the marker location information from the marker location filter 312 to convert that information into an estimated object pose. The object orientation filter 314 can be configured to then pass the estimated object pose information off to the motion database 310 and/or the controller interface 308 for use by the scanner controller 106.
(101) The tracking engine 304 can utilize various processes or algorithms in determining the location of marker reference points and converting this information into estimated object poses. Some examples of these processes or algorithms are described in more detail below. However, various other processes or algorithms can be used with the techniques disclosed herein.
(102) The controller interface 308 can be configured to convert object pose information into a coordinate system of the imaging scanner and to pass this information to the scanner controller 106. The controller interface 308 comprises a coordinate system converter 322 and a motion data feed generator 324. The coordinate system converter 322 can be configured to take the estimated object pose information from the tracking engine 304 and/or the motion database 310 and convert that object pose information from the motion tracking system's coordinate system into the scanner's coordinate system. The motion data feed generator 324 can be configured to take the converted object pose information and transmit it to the scanner controller 106. In some embodiments, the motion data feed generator 324 is configured to transmit object pose information to the scanner controller 106 immediately as the object pose information becomes available. In other embodiments, the motion data feed generator 324 can be configured to sync the timing of the motion tracking system 102 with the scanner controller 106. For example, the motion tracking system 102 may be configured to acquire images and estimate object poses at a rate of approximately 100 hertz. The scanner controller 106 and scanner 104 may, on the other hand, be configured to take scans at a different rate. The motion data feed generator 324 can therefore be configured to match the motion tracking system's output speed to the scanner and/or scanner controller speed. In some embodiments, this may involve caching converted object pose information until it is necessary to send that information to the scanner controller 106. In some embodiments, the motion data feed generator 324 is configured to obtain multiple object poses and to combine these, such as by averaging, before sending them to the scanner controller 106. In other embodiments, the motion data feed generator 324 is configured to transmit only the latest object pose estimate to the scanner controller 106. In some embodiments, the motion data feed generator 324 is configured to retransmit the last object pose estimate sent to the scanner controller 106 if the motion tracking system 102 has not generated a new object pose estimate since the last time the motion data feed generator 324 sent an object pose estimate to the scanner controller 106.
(103) As described above, although the motion compensation system 300 illustrated in
(104) Although the motion compensation system 300 illustrated in
(105)
(106) The reference point filter 336 can be configured to, among other things, analyze two dimensional images of an optical marker to determine the locations of the reference points of that marker. The principal quantities generator 338 can be configured to analyze the reference shape formed by the reference points, as viewed by two or more detectors, to determine a number of principal quantities, such as six, that can be used to help describe or define the 3D position and orientation of the optical marker.
(107) The reference frame translator 340 can be configured to convert between the two dimensional reference frame of each detector and the three dimensional frame of the motion tracking system. The error filter 342 can be configured to analyze differences in principal quantities based on the visualized reference points and based on estimates of an object pose to determine an amount of error between the two. The convergence filter 344 can be configured to perform iterative processes to reduce an amount of error in an object pose estimate until an object pose estimate has an acceptable amount of error. The tracking engine 334 can be configured to communicate with the motion database 310, calibration engine 306, and controller interface 308 similarly to as described above with respect to the motion compensation system 300 of
(108) Other Embodiments of Motion Compensation Systems
(109)
(110)
(111)
(112) Although the motion compensation system 440 comprises mirrors to redirect the lines of sight, other methods of redirecting a line of sight may be used, alone or in combination with mirrors. For example, fiber optics or prisms may be used to redirect a line of sight and create a virtual scissor angle.
(113)
(114) Optical markers may also be positioned at locations that are not rigid or substantially rigid. For example, an optical marker may be attached to a patient's skin. In some embodiments, such as when the marker is attached to a patient's skin, due to skin movement or skin elasticity, the marker may at times move in relation to the object being scanned, which can introduce inaccuracies into a medical imaging scan. Accordingly, in some embodiments, a motion compensation system can be configured to differentiate between movements of the object being scanned, such as a patient's head, and skin movement, which may not correlate to actual movement of the object being scanned. In some embodiments, the system can be configured to compare the positioning of two or more markers relative to themselves in order to differentiate between head movement and skin movement.
(115) Utilizing multiple optical markers 110 can have various benefits. For example, multiple markers may be used for redundancy, in case one or more markers is not currently visible to one or more detectors based on the current object's pose. Another advantage is that multiple optical markers can be analyzed simultaneously by the motion tracking system 102 to obtain multiple object pose estimates. Those multiple object pose estimates can then be combined to generate a single more accurate estimate. For example, the multiple estimates can be averaged to come up with an average estimate. In another example, there may be a measure of margin of error for each estimate and the estimates may be combined using a weighted average based on the margin of error. In other embodiments, only the most accurate estimate is used and other estimates are dropped.
(116)
(117) The motion compensation system 470 illustrated in
(118)
(119)
(120) The embodiment illustrated in
(121) The embodiment illustrated in
(122) Mounting of Optical Markers
(123) Optical markers can be mounted to an object being tracked in various ways.
(124) In some embodiments, the mounting portion 514 is configured to be shaped to a specific user to enhance the fit and/or rigidity of the optical marker 110. For example, the mounting portion 514 can be configured to be softened and placed onto the user's teeth and then hardened. In one embodiment, the mounting portion 514 comprises a polymer that softens when heated and hardens when cooled. In that example, the mounting portion 514 may, for example, be heated in warm water and then placed in the user's mouth to set. In another embodiment, the mounting portion 514 comprises an ultraviolet curable polymer. For example, the mounting portion 514 can be placed on a user's teeth in a soft condition and then cured to a hardened condition using an ultraviolet light. In other embodiments, the mounting portion 514 can be configured to have an adhesive or moldable polymer material placed into the trough 520 to conform to or adhere to a user's teeth.
(125) The mounting portion 514 can be connected to the optical marker 110 using the connection portion 522. In some embodiments, the optical marker 110 and mounting portion 514 are an integral inseparable unit. In other embodiments, the connection portion 522 enables the optical marker 110 to be detached and reattached to the mounting portion 514. This may be advantageous, for example, to enable the mounting portion 514 to be shaped to a user's teeth without the optical marker 110 mounted to it. In some embodiments, the connection portion 522 is relatively small and enables the optical marker 110 to be relatively close to the mounting portion 514. In other embodiments, the connecting portion 522 is longer and enables the optical marker 110 to be positioned further away from the object being tracked. In some embodiments, the connection portion 522 is configured to position the optical marker 110 such that it will not be covered by a patient's lips when the mounting portion 514 is positioned over the user's teeth. In some embodiments, the mounting portion 514 comprises an additional flap configured to be positioned over a user's lip to keep their lip from blocking the optical marker 110.
(126) In some embodiments, an optical marker can be internally lighted. For example, an optical marker can include one or more LED lights or the like inside the optical marker to uniformly or substantially uniformly light the optical marker or individual reference points. Illumination of an optical marker can be advantageous to enable a detector to more clearly image the optical marker and to make it easier for a motion tracking system to distinguish the reference points on the optical marker and/or to distinguish centroids of the reference points. In some embodiments, the detectors and/or another part of a motion compensation system includes lights configured to illuminate the optical marker. In some embodiments, such as when infrared detectors are used, the lighting can be infrared lights. In some embodiments, an optical marker can be configured to glow in the dark to enable enhanced illumination without using auxiliary lighting. In some embodiments, an optical marker can be configured to be charged with light by being placed in front of a light source for a period of time, after which the optical marker will glow for a period of time.
(127)
(128)
(129)
(130) The concept of utilizing distributed reference points may take various forms. For example, each reference point may be positioned on a block, substrate, or other backing, with the backing being affixed or attached to the patient or object being tracked. For example, the reference point 612 illustrated in
(131) Although the embodiments illustrated in
(132) In some embodiments of a motion compensation system, obstructions may exist that could potentially block a line of sight to an optical marker. For example, a head cage or head coil, such as is shown in
(133)
(134) In some embodiments, a motion tracking system can be configured to track the larger marker 610 and/or the smaller marker 110 based on the motion the detectors are capturing. For example, if a system is concerned with local motion or twitching of the patient's forehead skin that may introduce inaccuracies into the system, the system may be configured to stop tracking utilizing the larger optical marker 610 when motion in a direction consistent with a localized motion of a marker 614 is detected. For example, if a marker 614 is imaged moving in a longitudinal direction, such as up and down with respect to the patient's head, this may be indicative of local skin motion and may introduce inaccuracies if tracking the optical marker 610 is used. However, if motion of the reference point 614 in a side-to-side direction is detected, especially if both reference points 614 are moving in the same direction, this is less likely indicative of local skin motion and is more likely that the patient is turning his or her head. Accordingly, tracking based on the larger optical marker 610 may be more accurate during that motion than tracking the smaller optical marker 110. In one embodiment, a motion tracking system can be configured to track motion utilizing a larger distributed marker when motions not indicative of localized skin motion are detected, but to track using a smaller non-distributed marker when motions indicative of localized skin motion are detected.
(135)
(136) Motion Compensation System Processes
(137)
(138) At block 712, a controller interface generates motion data. For example, the controller interface 308 illustrated in
(139)
(140) At block 728, a tracking engine acquires simultaneous marker images from two different vantage points. For example, the tracking engine 304 illustrated in
(141) At block 732, an object orientation filter analyzes the reference point positions to determine a three dimensional object orientation or pose. For example, the object orientation filter 314 can be configured to analyze the information created at block 730 to estimate a three dimensional object pose in 6 degrees of freedom and, in some embodiments, to perform an iterative process to obtain a better estimate. The objection orientation filter can be configured to store this 3D object orientation or pose information in the motion database illustrated at block 734. At block 736, the controller interface can be configured to generate motion data. For example, the controller interface 308 illustrated in
(142) At block 742, the process flow varies depending on whether the imaging scan is complete. If the imaging scan is not complete, the process flow proceeds back to bock 728 and proceeds as described above. If the imaging scan is complete, the process flow proceeds to block 744 and the motion compensation process is completed.
(143)
(144) At block 760, a reference point filter locates reference points and the first two dimensional image. For example, the reference point filter 336 can be configured to analyze the first two dimensional image to locate the alternating ellipses or circles of the optical marker illustrated in
(145) At block 764, a principal quantities generator calculates baseline principal quantities or attributes based on the two dimensional reference point locations. For example, the principal quantities generator 338 illustrated in
(146) At block 766, a reference frame translator accesses an estimated object pose. For example, the reference frame translator 340 illustrated in
(147) At block 770, a reference frame translator calculates expected locations of reference points based on the estimated object pose. For example, the reference frame translator 340 illustrated in
(148) At block 774, an error filter determines an amount of error between the baseline and comparison principal quantities. For example, the error filter 342 illustrated in
(149) Returning to block 776, if the error is within an acceptable range, the process flow proceeds to block 780. At block 780, the tracking engine stores the estimated object pose in the motion database. At block 782, a controller interface generates motion data. For example, the controller interface 308 illustrated in
(150) At block 788, the process flow varies depending on whether the imagine scan is complete. If the imaging scan is complete, the process flow proceeds to block 790 and the motion compensation process is completed. If the imaging scan is not complete at block 788, the process flow proceeds back to block 758 and proceeds as described above.
(151) Although various embodiments described herein describe images from multiple detectors being acquired simultaneously or substantially simultaneously, in some embodiments, the images must merely be acquired within a certain time of each other. For example, in some embodiments, the images must just be acquired within a timeframe that is equal to or less than the frequency of the motion tracking system providing motion updates to the scanner controller. Another example of a process flow diagram illustrating an example of a motion compensation process is shown in
(152) In some embodiments, the processes illustrated in
(153)
(154) At block 709, a marker location filter analyzes images to determine two dimensional positions of reference points for the larger reference shape. For example, a marker location filter can be configured to determine the reference point locations of the reference points 614 and 612 illustrated in
(155) At block 713, the marker location filter analyzes the images from the detectors to determine 2D positions of reference points for the smaller reference shape. For example, the marker location filter can be configured to determine the positions of the reference points 612 and 616 illustrated in
(156) At block 717, the tracking engine combines both object orientation or pose estimates from blocks 711 and 715 to produce a single estimate. For example, the tracking engine may average the estimates. In another example, the tracking engine may use one estimate or the other depending on which is most likely to be the more accurate estimate at this point in time. The tracking engine can be configured to communicate with the motion database illustrated at block 719 to store the estimate. At block 721, a controller interface generates motion data. For example, the controller interface can be configured to convert the object orientation or pose estimate into a scanner's coordinate system. At block 723, the controller interface transmits the motion data to a scanner controller. At block 725, the scanner controller adjusts the scanner to compensate for the motion. At block 727, the process flow varies depending on whether the imaging scan is complete. If the imaging scan is not complete, the process flow proceeds back to block 707 and proceeds as described above. If the imaging scan is complete at block 727, the process flow proceeds to block 729 and the motion compensation process is completed.
(157) Optical Target Fixed to an Anatomical Location, e.g., the Head
(158) A challenge in optical head tracking is locating a head feature which moves rigidly with the body's skeletal frame. The skin is elastic and allows significant motion (relative to the displacement desired accuracy; for instance while blinking, twitching or wrinkling the nose or forehead. To overcome this challenge, in an embodiment, the system is configured to employ two or more optical tracking targets for placement on the patient. For example, two or more optical tracking targets can be coupled (for example, painted or affixed to the face of the patient) to the skin of the patient. By employing two or more optical tracking targets, the system can be configured to compensate for the elastic nature of the skin in order to determine the motion of the patient. For example, the system can be configured to track motion of the two or more optical tracking targets and average the detected motion in order to determine the approximate motion of the patient. Alternatively, in an embodiment, the system can be configured to analyze the detected motion from the two or more optical tracking targets and compare the detected motion from each to a predicted motion value. The system can be configured to select the detected motion value that is closest to the predicted motion value and ignore the rest of the detected values. Alternatively, the system can be configured to ignore the detected values that are substantially different from the predicted motion value. The system can be configured to average the detected motion values that have not been ignored. In an embodiment, the system can be configured to apply or combine one or more of the foregoing techniques.
(159) To overcome the challenge of the elastic nature of skin, in an embodiment, an optical tracking target can be coupled to the upper teeth of the patient. One accessible feature which is rigid to the skull is the upper teeth. Unlike the teeth on the lower jawbone, the upper teeth are rigidly affixed to the skull of the patient. In an embodiment, a compact and reliable optical tracking target can be attached to one or more of the upper teeth with a clip-on or other coupling device. Such attachment devices can be configured to be extremely comfortable. In an embodiment, a printed precision optical target is attached to the top front teeth of a patient.
(160) An optical target can be configured to be easy to locate with a high degree of accuracy regardless of orientation in a sensor field of view. A circle or series of concentric circles or ellipses can be potentially advantageous in this regard. Furthermore, to accommodate the fastest composite 2D data processing methods, a number (at least 3) of centroid positions can be discernible at every instant in time. The target can be, in some embodiments, composed of three sets of concentric circles or ellipses located at the vertices of an equilateral triangle. Compactness is desired for practical reasons, but the minimum size and spacing of the targets is dictated to large extent by characteristics of the sensors and the available non-occluded optical lines of sight through the MRI field compensation coil. A tradeoff arises, for instance, between the minimum size of the target and the cost of the imaging cameras used to sense the head motion—the smaller the edge dimension of the target triangle, the more pixels required in the camera sensor, and the faster the readout and processing electronics required.
(161) As a reasonable compromise, in some embodiments an equilateral triangle side length of 0.5 inches can be adopted. The printed target pattern includes a solid central elliptical dot of 1/16″ minor diameter at each triangle vertex, and each dot is surrounded by a first concentric ellipse of 3/16″ minor diameter and 1/32″ line width, and a second concentric ellipse of 5/16″ minor diameter and 1/32″ line width (ellipses scaled to look circular from camera nominal 45° look angle). In this embodiment, the entire target measures about 1 inch wide by about 0.93 inches high. Other dimensions are possible.
(162) A camera viewing this target is able to determine the centroid of each ellipse on the target pattern using a simple brightness moment calculation, independent of orientation of the target. The target itself subtends only a small portion of the camera field of view, but is recognizable by its high contrast and lack of gray scale. In embodiments the computer processor is programmed to track each of the three sub-targets by enclosing each of the three sub-targets within a sub-pixel array of 48×48 pixels and to calculate centroids of each sub-target by dividing (a) the sum of the product of pixel darkness and pixel position by (b) the sum of the pixel darkness of all of the pixels in the 48×48 sub-pixel array. The processor is also programmed to move each of the 48×48 pixel arrays so that its target is always located fully within the sub-pixel array. With sufficient camera spatial and brightness resolution and target illumination and contrast, centroid positional accuracy of about 0.1 pixels in row and/or column or less is achievable using this target.
(163)
(164) Latency
(165) Latency in the measurement of head motion using optical tracking techniques is comprised of the camera sensor integration and readout time, the target centroid determination time and the 6-DOF decomposition time. In order to reliably track head motions as fast as 2 cm/second and head rotations as fast as 10 degrees per second, a camera frame rate of about 100 Hz is desired, with electronic shuttering to freeze motion at rates up to 10 times this speed for sharp resolution of the optical target without blurring. A significant field of view is required to accommodate large motions, so fast camera readout without expensive mechanical tracking capabilities will require either a low pixel density or a camera with a larger focal plane but the ability to window a smaller region of interest for readout. Centroid and 6-DOF decomposition algorithms running in composite 2D, rather than full 3D space, and utilizing rapidly converging solution methods can be capable of returning solutions to the compensating head coil electronics at 100 solutions per second, with about 10 ms of latency. In some embodiments, the system can be configured to operate with a latency that enables it to update the scanner in between each image acquisition
(166) Cameras
(167) For a subject wearing or coupled with the optical head tracking target, the target size and subject rotation angles and translation position determine the physical location of the three target centroids precisely in three dimensions. With precise knowledge of these angles and the optical sensor (camera and lens) parameters—pixel pitch, lens focal length and radial distortion, camera location and orientation relative to nominal target position—the location of the target centroid projections on the focal plane sensor can be predicted to any level of accuracy even prior to measurement.
(168) In principle, the inverse problem should be equally simple as long as the 3D position of the target centroids can be ascertained optically. Using two cameras, a stereo view of the centroid projections can be used to determine the 3D location in space of each target centroid, and the 6-DOF displacement vector can then be determined through a simple matrix inversion. In practice, however, this approach leads to expensive and complicated requirements on camera pixel density, pixel count, camera alignment and camera calibration.
(169) An alternate unfolding approach dispenses with stereo ranging but uses separate 2D projections from two cameras without attempting to correlate absolute target positions on the two cameras. This approach eliminates the strict requirements on camera alignment and magnification matching characteristic of the stereo vision approach, and also relaxes the pixel density and count requirements needed to obtain the required positional accuracy (about 0.1 mm in translation and about 0.1 degrees in rotation) by about a factor of 20, resulting in significant savings in cost and processing speed. Although various embodiments described herein may not utilize stereo ranging, imaging, or vision, some embodiments may utilize concepts of stereo ranging, imaging, or vision, either alone or in combination with other object orientation determination techniques described herein.
(170) Even for this 2D measurement approach some basic steps can be taken to calibrate camera parameters once the cameras are integrated with the head coil; these can be performed at the manufacturing facility. These include measuring the projected pixel location of a single reference point on both cameras, as well as the camera magnification factors for pixel displacement per degree of rotation in pitch, yaw and roll, and per mm of translation along x, y and z. However, as stated before, it is not necessary that the cameras be exactly aligned in space (e.g. perfectly normal) or that their magnifications (lens focal length and distance to reference point) be identical, as is easily verified by simulation.
(171) Stereo Versus Composite 2D Vision Requirements
(172) With a single camera viewing the target from 45 degrees off of vertical in the target plane, the camera sees very little centroid displacement when the target moves in the direction of the camera (e.g. upward vertical translation equal to horizontal translation in the camera direction, with no rotation). Assuming a 7 micron pixel pitch, a 25 mm lens, and a working distance of 14 inches, target displacement in the camera direction may be at least 0.6 mm before the target can be detected as a 0.1-pixel increase in target centroid separation. However, as shown in
(173) Camera Depth of Field
(174) To accommodate head roll of +/−15 degrees plus the 0.85-inch target width at a working distance of 14 inches, the lens can be configured to provide sharp focus for distances between 13″ and 15.5″. At f/22, assuming a circle of confusion slightly smaller than a camera pixel (7 microns), a 25 mm focal-length lens provides this necessary depth of field a nominal 14-inch focus. At this working distance, the optical path can be folded with a turning mirror (
(175) In some embodiments, one possible camera that can be utilized or modified for use with systems and methods as disclosed herein, is produced by Allied Vision Technologies and designated the Prosilica GE-680 Monochrome CCD Camera. This camera features a Kodak KAI-0340 ⅓″ 640×480 VGA focal plane sensor with 7.4 μm square pixels and a fast Gigabit Ethernet output delivering up to 205 frames per second at 12-bit pixel depth. An inexpensive possible lens for use is an Edmund Optics TechSpec 25 mm high-resolution fixed focal length lens.
(176) For this camera and lens, at 14 inches from the target at 45° incidence, the 5/16″ diameter target circles project to ellipses on the camera, with the minor diameter of the largest ellipses at about 28 pixels and the major diameter at about 40 pixels. With sufficient S/N ratio (target illumination) and lens MTF (sharpness), this pattern should allow accurate centroiding to about 0.1 pixels in row and/or column or less. The entire projected target subtends about 128 H×168 V pixels, and allowing for head roll of +/−11.5 degrees, a camera with 640 horizontal pixels (pixel columns) can accommodate the entire field of interest without mechanical tracking provisions.
(177)
(178) Six Degree-of-Freedom Measurement and Reporting Algorithm
(179) In some embodiments, the MRI Head Tracker takes real-time input from two 2D imaging sensors and analyzes these data to determine and report motions in six degrees of freedom with minimal latency. This task can be performed by detecting and measuring the three centroid positions on the target and utilizing those positions with a reporting algorithm to determine the position of the patient's head.
(180) Six-Degree-of-Freedom Coordinate System
(181) In an embodiment, the system is configured to use a coordinate system for reporting 6-DOF motions to the MRI field compensation system that is a Cartesian system aligned with the symmetry axis of the head coil as shown in
(182) Coordinate definitions are adopted by the same conventions used in defining aircraft motion, except that the rotation directions are taken to be right-handed (positive for counter-clockwise rotation about the basis direction vectors):
(183) x is the longitudinal (chin-to-crown) direction, with values increasing toward the top of the head
(184) y is the transverse (left-to-right) direction, with increasing values toward the patient's right ear
(185) z is the up-down direction, with increasing values toward the ceiling
(186)
is the yaw angle or right-handed rotation about the z-axis (head lean toward shoulder while facing forward, zero at normal “square” position, positive values for patient leaning toward patient's right shoulder)
(187)
is the pitch angle or right-handed rotation about the y-axis (nodding “yes,” zero at normal “square” position, positive values for patient looking “upward”)
(188)
is the roll angle or right-handed rotation about the x-axis (shaking the head “no,” zero at normal “square” position, positive values for patient looking toward patient's left side).
(189) The origin of coordinates and angle zero references are arbitrary, as only relative motions are reported, however two convenient reference origin positions exist: 1) at the center of the target in its normal (“square”) head position, and 2) at the base of the neck at the point where the head swivels for nod, turn and lean motions. The latter is adopted here (as shown in
(190) Target Displacement Equations
(191) The full 6-DOF translation is composed of a 3-D displacement as well as a 3-axis rotation. To first order we assume that the skull moves as a rigid body about a single rotation point somewhere in the neck. From this point the translation becomes separable from the rotation, so this is chosen as the coordinate origin. The rotations are separated into roll, pitch and yaw as described above, and the translated position through rotation follows the Euler rotation matrix formulation as follows (using right-handed angle conventions). The x, y, and z displacement coordinates then follow the independent translations:
(192)
(193) Decomposing the six independent translations from the absolute and relative displacements of the measured target centroids is the subject of this effort. The 2D inverse problem is somewhat more difficult than the 3D problem, in that after the target centroid projections in focal plane row and column are determined, significant degeneracies remain in the unfolding matrices for each camera. Combining the data from both cameras removes these degeneracies through a series of interrelated, nonlinear equations. The fastest procedure for solving this inverse problem is obtained by the Newton-Raphson method or a variant thereof, whereby an approximate first-order solution is proposed and tested against the known (measured) centroid locations on the two camera focal planes. The residual error is divided by the local derivatives with respect to each component of rotation and translation, to determine an iterative correction. The first-order solution is chosen by considering the features of the projected target pattern which are most strongly affected by a single rotation angle or displacement, and linearizing the inversion problem along these feature axes.
(194) A 6-DOF motion simulation and decomposition algorithm was developed and tested to allow simulation of arbitrary motions and then verify the ability of a pair of orthogonal cameras to decompose centroid measurements at the 0.1-pixel level into distinct x, y, z, roll, pitch and yaw components at the requisite level of accuracy.
(195) Six-Degree-of-Freedom Motion Determination Algorithm
(196) General subject motion is a superposition of translation along x, y, and z as well as rotation about the x, y and z axes (designated roll, pitch and yaw respectively). Displacements along each of these degrees of freedom are not sensitive to coordinate system origin; however it is convenient (as explained above) for modeling purposes to place an origin near the region of the spine about which the head rotates and swivels, and a secondary reference point at the center of the optical tracking target in the nominal (“correct”) head position and orientation. This secondary reference is typically offset from the spinal origin by ˜10 cm in x and ˜10 cm in z.
(197) The target shown in
(198) Yaw
(199) Rotation about the z-axis is designated as yaw; a positive or “right handed” rotation about this axis (head leaning to subject's right shoulder) results in a counterclockwise rotation of the target. Because this rotation usually occurs about a point lower in the neck, it is typically accompanied by a translation to the subject's right side (camera left), as seen in
(200) The median of the centered target triangle (as shown at the right in
(201) Pitch
(202) Rotation about the y-axis is designated as pitch; a positive or “right-handed” rotation about this axis (head tipped back) results in motion of the target upward off the gantry (+z) and toward the top of the head (+x). For a single camera this projection is not easily distinguishable from a simultaneous target displacement in x and y (see
(203)
(204) Roll
(205) Rotation about the x-axis is designated as roll; a positive or “right-handed” rotation about this axis (head pointing toward subject's left side) results in a motion of the target toward the subject's left (−y). For a single camera this motion is not easily distinguishable from a displacement in y (see
(206) As shown in
(207) X-Axis Translation
(208) Translation along the x-axis (positive toward top of head) results in a motion of the target along the vertical direction of the camera focal plane (see
(209) Y-Axis Translation
(210) Translation along the y-axis (positive toward subject's right side) results in a motion of the target along the horizontal axis of the camera focal plane (see
(211) Z-Axis Translation
(212) Translation along the z-axis (positive toward the ceiling) results in apparent motion of the target along the horizontal axis of the camera focal plane. Unlike for y translation, however, the direction of the horizontal displacement is opposite between the left-side and right-side cameras (see
Non-Degenerate Target Motion Parameters
(213) Pitch Versus (X+Z) Translation Degeneracy
(214) Pitch is nearly degenerate with simultaneous x and z translation, except for a small tilt in the triangle vertical which results from the tilt of the target plane about the y axis. This tilt creates an apparent clockwise rotation of the triangle from the left-side view and an apparent counterclockwise rotation from the right side view, as shown in
(215) Roll Versus (Y+Z) Translation Degeneracy
(216) Roll is nearly degenerate with simultaneous y and z translation, except for larger camera-to-camera differences in apparent target size encountered with roll, resulting from tilt of the target's plane about the x-axis. A significant difference in the apparent length of the target triangle base is a reliable distinguishing characteristic of roll motion rather than simple translation.
(217) For a roll of 0.1 degrees and translations in y and z of −0.244 mm and 0.0002 mm respectively, the lower centroid is unchanged in both camera views. In this case, the left-side camera sees the target triangle base shrink by 0.15 pixels while the right-side camera sees the triangle base grow by 0.15 pixels. Assuming the centroiding routine can locate the target centroids to an accuracy of 0.1 pixels, shifts of 0.14 pixels should be discernible, so a pitch as small as 0.1 degrees is distinguishable from a simple translation by comparison of the length of the target triangle base.
Six-Degree-of-Freedom Motion Determination Algorithm Architecture
(218) Complementary Projections Versus Stereo Imaging
(219) The target size, rotation angles and translation vector determine the relative displacement of the three target centroids precisely in three dimensions. Precise knowledge of camera and lens parameters (e.g., pixel pitch, lens focal length and radial distortion, camera location and orientation relative to nominal target position), are then sufficient to predict the location of the target centroid projections to better than 0.1 pixels in row and column for each camera. In principle, the inverse problem should be equally simple; the stereo view of the centroid projections determine the 3D location in space of each target centroid, and the 6-DOF displacement vector can then be determined through a simple matrix inversion. In practice, however, this approach leads to expensive and complicated requirements on camera pixel density, pixel count, camera alignment and camera calibration. An alternate unfolding approach dispenses with stereo ranging and uses the two camera projections separately without strict requirements on precise matching of camera alignment and magnification, to determine the 6-DOF displacement vector to within 0.1 degrees in each rotation angle and 0.1 mm along each translation axis. This approach relaxes the pixel density and count requirements by about a factor of 20 relative to the stereo approach, resulting in significant savings in cost and processing speed.
(220) Even for this 2D approach some basic measurements can be made to calibrate camera parameters once the cameras are integrated with the head coil; these can be easily performed at the manufacturing facility. These measurements include the projected pixel location of a single reference point on both cameras, as well as the camera magnification factors for pixel displacement per degree of rotation in pitch, yaw and roll, and per mm of translation along x, y and z. However, as stated before, it is not necessary that the cameras be exactly aligned in space (e.g. perfectly normal) or that their magnifications (lens focal length and distance to reference point) be identical, as has been easily verified by simulation.
(221) Inversion Equations
(222) The 2D inversion problem is somewhat more difficult than the 3D problem, in that after the target centroid projections in focal plane row and column are determined, significant degeneracies remain in the unfolding matrices for each camera. Combining the data from both cameras removes these degeneracies through a series of interrelated, nonlinear equations. The fastest procedure for solving this inverse problem is obtained by a variant of the Newton-Raphson method, whereby an approximate first-order solution is proposed and tested against the known (measured) centroid locations on the two camera focal planes. The residual error is divided by the local derivatives with respect to each component of rotation and translation, to determine an iterative correction. The first-order solution is chosen by considering the features of the projected target pattern which are most strongly affected by a single rotation angle or displacement, and linearizing the inversion problem along these feature axes.
(223) 6-DOF Extraction Algorithm
(224) Below is described one embodiment of a method for extracting the 6 degree of freedom displacement matrix from the observed target location on two focal plane cameras. Other embodiments may be used consistent with the disclosure herein. Further, although the embodiment below refers to steps, in some embodiments not all steps are included, additional steps are included, and/or the steps are not always performed in a set order.
(225) Step 1: Characterizing the Target Images
(226) The optical target consists of elliptical targets shown in
(227)
where is X.sub.0i,j the initial (zero displacement) horizontal camera coordinate of each centroid projection. b) Δ.sub.HD—the difference between the horizontal displacements (in pixels) of the target center for camera 1 and camera 2; the formula used is
(228)
(229)
where Y.sub.0i,j is the initial (zero displacement) vertical camera coordinate of each centroid projection. d) Δ.sub.BL—the difference in the apparent base length of the target triangle (in pixels) for camera 1 and camera 2; the formula used is
(230)
(231)
(232)
Step 2: Characterizing Global Variation in Principal Quantities with 6-DOF Motions
(233) Partial derivatives relative to subject displacements and rotations (φ, θ, ψ, Δx, Δy, Δz), of the principal quantities described above, about the initial (non-displaced) position, are computed numerically. Here: Roll φ is right-handed rotation about the x-axis Pitch θ is right-handed rotation about the y-axis Yaw ψ is right-handed rotation about the z-axis Δx is toe-to-head direction Δy is left-to-right direction Δz is down-to-up direction
(234) Starting from an initial target position in 3-D world space, defined as (φ, θ, ψ, Δx, Δy, Δz)=(0, 0, 0, 0, 0, 0), the initial target vertex world coordinates (x.sub.0i,y.sub.0i,z.sub.0i) are determined for vertex index i=1 to 3, based on the geometric size and shape of the target triangle and definition of a convenient coordinate origin.
(235) Local partial derivatives of each of the principal quantities, with respect to each of the 6 degrees of freedom (roll, pitch, yaw, dx, dy, dz), are performed numerically by evaluating changes in these quantities for small increments in each degree of freedom. Changes in the target vertex positions for specified motions along the six degrees of freedom are computed using the Euler rotation matrices and translation vector:
(236)
(237) Subsequently, the camera projections of these new target vertex positions are determined using a geometric projection calculation. Given accurate knowledge of camera positions, pixel size and lens focal length, the horizontal and vertical pixel numbers on each camera focal plane (camera index j equal to 1 or 2) that these new 3-D positions in space should project onto is as follows:
(238)
(239) Here and X.sub.i,j are Y.sub.i,j the horizontal and vertical pixel numbers for translated target vertex i projected onto the camera j sensor, X.sub.0,j and Y.sub.0,j are the horizontal and vertical number of the pixel column and row intersected by the optical axis of that camera (typically at or very near the camera center), f.l. and s.sub.pix are the lens focal length and camera pixel pitch, and the angles α.sub.i,j and β.sub.i,j are the polar and azimuth angles locating target vertex i, relative to the camera j focal axis. These angles are calculated from the vertex world coordinates as follows:
(240)
where the point (x.sub.⊥,j,y.sub.⊥,j,z.sub.⊥,j) is the point of intersection between the camera optical axis and the plane perpendicular to the optical axis which includes the translated target vertex (x.sub.⊥i,y.sub.⊥i,z.sub.⊥i):
x.sub.⊥,j=x.sub.0+κ(x.sub.cj−x.sub.0);y.sub.⊥,j=y.sub.0+κ(y.sub.cj−y.sub.0);z.sub.⊥,j=z.sub.0+κ(z.sub.cj−z.sub.0), [5]
with (x.sub.⊥cj,y.sub.⊥cj,z.sub.⊥cj) defining the 3-D position of camera j, (x.sub.⊥0,y.sub.⊥0,z.sub.⊥0) defining the nominal boresight position of both cameras at the un-displaced target center and the constant κ based on geometric projection and given by:
(241)
(242) In equation [4], the inverse cosine function is taken to range from 0 to π, and the appropriate sign for β.sub.i,j is given by:
sign[β.sub.i,j]=sign[(z.sub.cj−z.sub.i){(x.sub.cj−x.sub.0)(x.sub.⊥,j−x.sub.i)+(y.sub.cj−y.sub.0)(y.sub.⊥,j−y.sub.i)}−(z.sub.⊥,j−z.sub.i){(x.sub.cj−x.sub.0).sup.2+(y.sub.cj−y.sub.0).sup.2}]
(243) During this determination of the camera projection of the 3-D target vertices, a compensation function may be applied for large values of the polar angle α.sub.i,j to account for barrel distortion in the lens, based on prior lens calibration measurements. The geometric value for α.sub.i,j is first computed based on equation [3] and then adjusted for lens distortion by way of a pre-determined look-up table or measured fit function, and this new compensated value for α.sub.i,j is then used in the calculation of X.sub.i,j and Y.sub.i,j through equation [2].
(244) To numerically evaluate the partial derivatives of the principal quantities about the initialized target position, the un-displaced 3-D target vertex coordinates (x.sub.0i,y.sub.0i,z.sub.0i) are first projected to camera coordinates using equations [2] through [6] above, and initial values are computed for each of the principal quantities described in Step 1 (most should be zero or near-zero at the starting position). Then small increments of roll, pitch, yaw, x-, y- and z-axis displacements are introduced one at a time; for each increment the new world coordinates and the new camera projections of the target vertices are computed and the principal quantities are re-calculated. The change in each principal quantity is divided by the small angular or displacement increment to determine the partial derivative.
(245) For instance, to determine the partial derivatives with respect to roll, the displacement vector (φ, θ, ψ, Δx, Δy, Δz)=(ôφ, 0, 0, 0, 0, 0) is introduced to the general displacement equation [1] to determine the translated target vertex positions (x.sub.i,y.sub.i,z.sub.i). The conversion to camera coordinates is then performed using equations [2] through [6], and the principal quantities are calculated as outlined in Step 1. The difference between each principal quantity and the corresponding value of that quantity for the un-displaced calculation is divided by the small increment in roll, to give the partial derivative of each quantity with respect to roll. To determine partial derivatives with respect to pitch, the displacement vector (φ, θ, ψ, Δx, Δy, Δz)=(0, δθ, 0, 0, 0, 0) is used to initiate the calculations, and so on for all six degrees of freedom.
(246) Each of these six repetitions produces one column of the global partial derivative matrix:
(247)
Step 3: Determining First-Order Displacement Vector
(248) A first-order approximation to the displacement matrix is determined by multiplying the matrix of measured principal quantities, as determined in Step 1, by the inverse of the partial derivative matrix computed in Step 2:
(249)
Step 4: Characterizing Local Variation in Principal Quantities with 6-DOF Motions
(250) First order values for (φ, θ, ψ, Δx, Δy, Δz) determined in Step 3 are entered into the translation equation [1] to determine the corresponding translated 3-D target position (x.sub.⊥i,y.sub.⊥i,z.sub.⊥i) for each of the three target vertices. These world coordinates are projected to camera coordinates (X.sub.i,j, Y.sub.i,j) using equations [2] through [6], and the principal quantities are re-calculated. These six quantities are compared against the measured values of these quantities determined in Step 1, to create a residual error matrix:
(σ.sub.Σ.sub.
(251) Local partial derivatives of the principal quantities are calculated by introducing small increments in roll, pitch, yaw, x-, y- and z-axis displacements one at a time as before, but this time the increments are relative to the first-order displacement vector. For each increment, the new world coordinates and the new camera projections of the target vertices are re-computed and the principal quantities are re-calculated. The change in each principal quantity is divided by the small angular or displacement increment to determine a local partial derivative. For instance, to calculate partial derivatives with respect to roll, the first-order displacement vector {φ.sub.0, θ.sub.0, ψ.sub.0, (Δx).sub.0, (Δy).sub.0, (Δz).sub.0} is replaced by {φ.sub.0+ôφ, θ.sub.0, ψ.sub.0, (Δx).sub.0, (Δy).sub.0, (Δz).sub.0} and resulting changes to each of the principal quantities is divided by ôφ to determine the local derivative with respect to roll. This is repeated for each of the six degrees of freedom.
(252) Each of these six repetitions produces one column of the new local partial derivative matrix:
(253)
Step 5: Determining Coarse Correction to First-Order Displacement Vector
(254) A coarse correction is computed to improve the first-order displacement vector and reduce residual error, by multiplying the residual error matrix determined in Step 4 by the inverse of the local partial derivative matrix, also determined in Step 4:
(255)
(256) The first-order displacement vector is incremented by the coarse correction matrix to create a better approximation to the displacement vector:
{φ.sub.0+Δφ,θ.sub.0+Δθ,ψ.sub.0+Δψ,(Δx).sub.0+Δ(Δx),(Δy).sub.0+Δ(Δy),(Δz).sub.0+Δ(Δz)}.
Step 6: Performing Fine Correction to Determine Final 6DOF Displacement Vector
(257) Steps 4 and 5 are repeated, starting with the coarse-corrected displacement vector, to determine a final fine correction to the displacement vector. After this iteration, the resultant fine correction increments are added to the coarse-corrected vector to create the final 6-DOF displacement vector. Empirical results from a general simulation indicate that this fine correction is sufficient in all cases to reduce residual errors to well below the stated 0.1-degree, 0.1-mm tolerances.
(258) Algorithm Numerical Simulation to Verify Absolute Convergence
(259) As cameras, targets and rotation stages are being procured and assembled, the 6 DOF decomposition algorithm can be coded and tested for a test set of rotations. It is clear that the routine will converge for small translations and rotations, but it can be potentially advantageous to determine whether there are limitations on its convergence for extreme displacements in all six degrees of freedom. To this end, we imagine an extreme target displacement, calculate the 3D position of the displaced target, calculate the centroid positions that will be seen on each of the two cameras, and run the decomposition algorithm to determine speed of convergence.
(260) In some embodiments, to demonstrate absolute convergence of the iterative 6DOF unfolding algorithm, the simulation is started with a test set of very large rotations and displacements, as listed in Table 1 below.
(261) TABLE-US-00001 TABLE 1 Example of a set of extreme angular rotations and linear translations of an imaginary patient for purposes of testing algorithm convergence. Head Yaw (Lean) Psi (deg toward patient's right shoulder) 8.0000 Head Pitch (Nod) Theta (deg relative to level; pos is toward 15.0000 top of head) Head Roll (Shake) Phi (deg relative to square; pos toward 12.0000 patient's left side) Head shift dx (mm toward top of head) −9.0000 Head shift dy (mm to patient's right) 3.0000 Head shift dz (mm away from table) 7.0000
(262) The simulation begins by determining the locations of the displaced centroids that will be seen by each camera, allowing for some degree of mispointing and misalignment of each camera. The original (nominal) target location is rotated and displaced by the Euler rotation formalism presented in Section 2.5.2.2, to determine the three displaced target centroid locations in three-dimensional space. Next these “world coordinates” are translated to 2-D “camera coordinates” for each of the two cameras independently, as described in the same Section.
(263) Assuming the target is imaged into these camera coordinates, but that the operator has no prior knowledge of the displacement matrix giving rise to this target position, we use the algorithm as described in Section 2.5.2 from end to end to recreate the displacement matrix. By the end of Step 3 (Section 2.5.2.3), the algorithm returns an initial estimate of the 6DOF displacement vector, as shown in Table 2 below.
(264) TABLE-US-00002 TABLE 2 First estimate of 6DOF displacement based on method described in Section 2.5.2. First Approximation Yaw (degrees) 4.4313 First Approximation Pitch (degrees) −19.4474 First Approximation Roll (degrees) 8.8784 First Approximation X displacement −6.4257 (mm) First Approximation Y displacement −2.5639 (mm) First Approximation Z displacement −5.9428 (mm)
(265) As expected, residual errors at this stage are atypically large, due to the extreme magnitudes of the translations and rotations chosen for this simulation along and about each axis; this situation creates a good test for absolute convergence of the Newton Raphson algorithm methodology. Assuming this estimate to be correct, the algorithm in Step 4 (Section 2.5.2.4) again calculates the displaced position of the target, the resulting centroid positions seen by each camera, and the principal quantities (vertical tip sum and difference, base length difference, vertical displacement sum, and horizontal displacement sum and difference) which would result, for comparison with the actual observed values. The residual errors, in pixels, and the local derivatives of each of the principal values for small changes (pixels per 0.1 degrees) in yaw, pitch, and roll, and for small changes (pixels per 0.1 mm) in dx, dy and dz are calculated as described in Section 2.5.2.4, and tabulated as shown in Table 3 below.
(266) TABLE-US-00003 TABLE 3 Residual Error (in pixels) and local derivatives with respect to Yaw, Pitch, Roll (pixels per 0.1 deg), x-displacement, y-displacement, and z-displacement (pixels per 0.1 mm), of the principal quantities Vertical Tip Sum, Vertical Tip Difference, Base Length Difference, Vertical Displacement Sum, Horizontal Displacement Sum, and Horizontal Displacement Difference. Residual ∂/∂Y ∂/∂P ∂/∂R ∂/∂x ∂/∂y ∂/∂z Error VT1 + VT2 0.2575 0.0383 −0.099 0.0021 0.0045 0.0021 −3.6538 VT1 − VT2 0.0657 −0.2756 −0.0131 0.0005 0.0018 0.0274 6.8709 BL1 − BL2 −0.5223 0.0277 0.4988 0.0109 −0.0702 0.0105 −2.9918 VDT + VD2 −0.3118 5.8134 0.0350 1.8813 0.0112 −0.2223 −168.5591 HD1 + HD2 −2.5875 −0.1680 3.3651 0.0117 −1.3090 −0.0124 58.1859 HD1 − HD2 0.5823 1.4452 0.7697 −0.0140 0.1114 −1.4280 20.793
(267) The matrix of derivatives at the left of Table 3 is inverted and multiplied by the residual error vector at the right, to yield first-order corrections to the initial estimate of the displacement vector, as described in Section 2.5.2.5, and as shown at the left of Table 4 below. These are added to the initial estimates, to produce the more refined estimate of the 6 DOF displacement vector, shown at the right of Table 4.
(268) TABLE-US-00004 TABLE 4 (left) First-Order Corrections to Initial Estimates of Yaw, Pitch, Roll, dx, dy and dz, obtained by inverting the matrix of derivatives at left of Table 3 above and multiplying this inverse matrix by the residual error vector at right of Table 3. These corrections are added to initial 6DOF motion estimates to produce improved estimates at right above. Yaw Adjustment 3.8632 First Newton Iteration Yaw 8.2945 (deg) (deg) Pitch Adjustment 4.5672 First Newton Iteration Pitch −14.8803 (deg) (deg) Roll Adjustment 3.5642 First Newton Iteration Roll 12.4426 (deg) (deg) dx Adjustment −3.0846 First Newton Iteration Delta −9.5103 (mm) X (mm) dy Adjustment 6.5969 First Newton Iteration Delta 4.0329 (mm) Y (mm) dz Adjustment 12.9426 First Newton Iteration Delta 6.9998 (mm) Z (mm)
(269) This process is repeated for a second and final time as described in Section 2.5.2.6, assuming again that the (now refined) 6 DOF displacement vector is accurate, and calculating first the 3D target centroid positions and then the locations of the target centroids as projected onto each of the two camera focal planes. Again the six principal quantities are computed and compared with the actual observations to produce a vector of residual errors. Again the local derivatives are computed, this time at the location of the first-order displacement vector. The results are tabulated as shown in Table 5 below.
(270) TABLE-US-00005 TABLE 5 First-Order Residual Error (in pixels) and new local derivatives with respect to Yaw, Pitch, Roll (pixels per 0.1 deg), x-displacement, y-displacement, and z-displacement (pixels per 0.1 mm), of the principal quantities Vertical Tip Sum, Vertical Tip Difference, Base Length Difference, Vertical Displacement Sum, Horizontal Displacement Sum, and Horizontal Displacement Difference. Residual ∂/∂Y ∂/∂P ∂/∂R ∂/∂x ∂/∂y ∂/∂z Error VT1 + VT2 0.2498 0.0545 0.0785 0.0020 0.0028 0.0007 0.1715 VT1 − VT2 0.0682 0.2935 0.0223 0.0012 0.0034 0.0212 0.0827 BL1 − BL2 −0.3146 0.0536 0.4966 0.0372 0.0723 0.0094 0.5096 VD1 + VD2 −0.5927 5.7797 0.0405 1.9353 0.0084 0.1911 −4.3941 HD1 + HD2 −2.5162 0.3237 3.7395 0.0074 1.3067 0.0135 4.8578 HD1 − HD2 −0.6876 1.779 0.7547 0.0177 −0.0884 −1.4784 2.5723
(271) The matrix of derivatives at the left of Table 5 is inverted and multiplied by the residual error vector at the right, to yield final corrections to the first-order estimate of the displacement vector, as shown at the left of Table 6 below. These corrections are added to the first-order estimates, to produce the final second-order estimate of the 6 DOF displacement vector, shown at the right of Table 6.
(272) TABLE-US-00006 TABLE 6 (left) Second-Order Corrections to First-Order Estimates of Yaw, Pitch, Roll, dx, dy and dz, obtained by inverting the matrix of derivatives at left of Table 5 above and multiplying this inverse matrix by the residual error vector at right Table 5. These corrections are added to first-order correction obtained by the same method, to produce final values for each of the 6 DOF motions used in the simulation. Yaw Adjustment −0.2947 Final Yaw 7.9999 (deg) (deg) Pitch Adjustment 0.2440 Final Pitch 15.0003 (deg) (deg) Roll Adjustment 0.4408 Final Roll 11.9978 (deg) (deg) dx Adjustment 0.5114 Final Delta X −8.9989 (mm) (mm) dy Adjustment 1.0377 Final Delta Y 2.9952 (mm) (mm) dz Adjustment 0.0058 Final Delta Z 6.994 (mm) (mm)
(273) Even for the extreme rotations and displacements used in this model, the algorithm is shown to converge to within 0.003 degrees and 0.006 mm in only two iterations. Given the number of floating-point operations needed to perform the initial estimate and two successive iterations of the Newton method, the algorithm can produce a solution on a typical laptop computer in less than 5 milliseconds.
(274) Quaternion Representation
(275) The head coil ICD specifies the rotation vector in terms of the quaternion, for which (still using right-handed Euler angle rotation conventions):
(276)
The translation vector is unchanged from the form calculated here.
Centroid Determination Algorithm
(277) The centroid location on the focal plane is given by:
(278)
(279) This calculation is performed for three subregions 2010 on the target as shown in
(280) Centroid Determination
(281) In some embodiments, a test target can be printed and mounted in the view field of a monochrome camera at an angle of approximately 45 degrees. At this angle the elliptical target projected to an approximately round target on the camera focal plane. In some embodiments the camera can be focused at a full-scale printed target oriented at 45 degrees at a distance of 14.1 inches. Camera field of view is roughly the size of the rectangle in the center of the camera calibration target mounted next to the target.
(282) The calculated target centroid is displayed as a red dot at the center of a LabView image, and displayed as a floating point (x,y) pair to the right of the image. At illumination levels above about 20% of full scale, the measured centroid location does not fluctuate above the 0.1-pixel level in row or column; for lower intensity levels, statistical fluctuations exceed this threshold. It is noted, however, that for the black-on-white printed target, uniformity of illumination can be potentially important—if the target is illuminated significantly more strongly from the left or right side, for instance, the moment calculation could add bias in the horizontal direction and would shift the centroid outside of the specified error threshold. This effect could in some cases put an undesirable cost constraint on the illumination approach, so an intensity thresholding algorithm is first implemented, by which the target histogram is clipped near the lower extrema for the bright and dark region intensities, eliminating the undesirable effect. In some embodiments, a Camera Control screen view can allow control of camera frame rate and readout resolution, showing manually-selected region of interest. Full camera field of view is approximately represented by a black region on the screen. The centroid can be displayed as a red dot at the center of the circular target, and camera x-y coordinates are displayed as floating point numbers to 2-decimal precision to the right of the display.
EXAMPLE 1
(283) Camera Calibration
(284) As with any camera lens, the lens used for the head tracker could have some level of distortion as a function of distance from imaging axis. Azimuthal distortion should be negligible, but radial distortion can be measured after lens installation and fit to a polynomial curve to allow rapid compensation of centroid positions near the edges of the camera field of view. The 6DOF unfolding algorithm can be constructed to accommodate typical levels of radial distortion as a second-order compensation during the application of the Newton Raphson iteration method.
(285) Radial distortion can be determined using a printed reference target with concentric circles of diameter ⅓″, ⅔″, 1″, and so on up to a maximum diameter of 4 inches, as shown in
(286) In one embodiment, the measured radial distortion measured for the TechSpec High Resolution Fixed Focus 25 mm lens follows camera polar angle θ.sub.C=(1+0.0053144θ−0.0016804θ.sup.2+0.0002483θ.sup.3−0.0000138θ.sup.4)θ, with laboratory polar angle θ in degrees. At the extreme corner of the viewing field, where θ˜6.75°, camera aberration results in a radial growth in camera angle of about 0.7% relative to true angle, or about 2.8 pixels in radius.
(287) Full 6-DOF Tracking
(288) The full 6-DOF tracking algorithm was coded in LabView with the Graphical User Interface (GUI). The upper left side of the GUI screen gives centroid information for target circles in the current frame, and the lower left side gives the same information for the prior frame. For each, one nested target circle from the set of three is displayed in negative (white on black) along with a histogram of its pixel brightness within a 48-by-48 pixel box centered on the centroid location of the previous frame. This histogram is split into two sections to display (at left) the peak from background pixels at one end of the brightness scale, and (at right) the peak from the pixels of the target itself, at the other end of the brightness scale. A long continuum of pixels in between represents pixels at dark-light boundaries in the target frame. From analysis of the two histograms, the target field is clipped at the lower-brightness shoulder on the bright side, and the upper brightness shoulder on the dark side, to create a binary target field that is not sensitive to variations in illumination across the target. Although displayed in real time for only one target circle, all three target circles are processed in this way.
(289) Next to the target histograms, the x-y camera centroid locations are displayed to two-decimal precision for each of the three nested circle targets; again at the upper half of the screen for the current data and at the lower half of the screen for the prior frame.
(290) The right side of the screen displays the processed 6-DOF data, after analysis using the approach described in Section 2.5. An analog meter-style display shows the acquisition and processing time per frame, which is limited at its low end to the camera frame integration and readout time of about 8 milliseconds. Using a single iteration of the Newton-Raphson routine described in Section 2.5, the algorithm runs during the integration period for the successive frame, so the processing time is approximately 8 milliseconds, corresponding to a 120 Hz camera readout rate. The 6-DOF data can be displayed in either analog or digital format, but the digital format can be read to precision of 0.01 mm and 0.01 degree for comparison with the 0.1 mm, 0.1 degree accuracy requirements.
(291) Laboratory Mechanical Layout for Head Tracking Simulation
(292) The laboratory setup was designed to mimic head rotation and displacement using a six-degree-of-freedom optical rotation mount. This mount included three ganged translation stages along the x-, y-, and z-axes of the optical table, and three ganged rotation stages corresponding to yaw, roll and pitch respectively. The two monochrome cameras and turning mirrors were mounted in the appropriate geometry for use with an existing 12-channel head coil. The two monochrome cameras are in foreground, mounted at ±45° relative to horizontal to accommodate rotation by the turning mirrors. The turning mirrors are mounted 10 inches behind cameras (slightly obscured by the cameras in the picture). The target is partially visible in the reflection of each mirror. The 6-DOF rotation stage is at center in foreground, with the y-axis stage at bottom, x-axis stage next, and z-axis stage above that, followed by the yaw rotation stage, the roll stage, and finally the pitch stage with target at the top (the pitch rotation handle is obscured by the stage). A near-IR illumination LED is at the center in background; light from this stage is within the camera spectral range, but hardly visible to the human eye.
(293) X-Axis Translation
(294) The second translation stage from the bottom in the 6-DOF displacement assembly controls x-axis displacement (aligned with the patient's spine). The x-axis translation stage control knob is turned four full rotations (corresponding to −2.54 mm), and the absolute position change is calculated from the resulting motion of the centroid camera coordinates. Results are: the displacement determined by the unfolding algorithm is −2.56 mm in x, less than 0.1 mm in y and z, and less than 0.1° in roll, pitch and yaw. The target displacement by dx=−2.54 mm, with zoom on lower right display section of GUI showed calculated dx=−2.56 mm, dy=0.08 mm, dz=0.02 mm, dφ=0.05°, dθ=−0.03°, and dψ=−0.01°.
(295) Y-Axis Translation
(296) The bottom translation stage in the 6-DOF displacement assembly controls y-axis displacement (patient's left-to-right). The y-axis translation stage control knob is turned four full rotations (corresponding to −2.54 mm), and the absolute position change is calculated from the resulting motion of the centroid camera coordinates. This resulted in a target displacement by dy=−2.54 mm, with zoom on lower right display section of GUI showing dx=0.00 mm, dy=−2.47 mm, dz=−0.01 mm, dφ=0.64°, dθ=−0.04°, and dψ=−0.03°.
(297) Z-Axis Translation
(298) The top translation stage in the 6-DOF displacement assembly controls z-axis displacement (patient's down to up, with the patient lying on his back). The z-axis translation stage control knob is turned four full rotations (corresponding to −2.54 cm), and the absolute position change is calculated from the resulting motion of the centroid camera coordinates. The displacement determined by the unfolding algorithm was −2.54 mm in z, less than 0.1 mm in x and y, and less than 0.1° in roll, pitch and yaw. The results were a target displacement by dz=−2.54 mm, with zoom on lower right display section of GUI showing dx=0.01 mm, dy=−0.01 mm, dz=−2.59 mm, dφ=−0.02°, dθ=−0.06° and dψ=0.01°.
(299) Yaw Rotation
(300) The bottom rotation stage in the 6-DOF displacement assembly controls yaw rotation (patient's left shoulder—to—right shoulder lean direction). The yaw rotation stage control knob is turned by +4° degrees (heading 315° to heading 311° on stage, corresponds to movement toward right shoulder), and the absolute position change is calculated from the resulting motion of the centroid camera coordinates. The displacement determined by the unfolding algorithm is less than 0.1 mm in dx, dy and dz, 0.1° in roll and less than 0.1° in pitch, and 3.94° in yaw. The results were a target rotation by dψ=+4.00°, with zoom on lower right display section of GUI showing dx=0.07 mm, dy=−0.05 mm, dz=0.02 mm, dφ=0.10°, dθ=−0.01°, and dψ=3.94°.
(301) Roll Rotation
(302) The middle rotation stage in the 6-DOF displacement assembly controls roll rotation (patient's right shoulde—to—left shoulder “head shaking” direction). The roll goniometer control knob is turned by +5° degrees, and the absolute position change is calculated from the resulting motion of the centroid camera coordinates. The displacement determined by the unfolding algorithm is less than 0.1 mm in dx, and dz, 1.78 mm in dy, 4.97° in roll and less than 0.1° in pitch and yaw. Displacement in y is expected due to the fact that the center of rotation for the Thorlabs GNL18 goniometer stage is 44.5 mm above the mount surface, while the target is only 22 mm above the stage. For the resulting −20.5 mm lever arm, the y-displacement due to a 5° roll rotation is −(−20.5 mm)*sin(5°)=+1.79 mm, in good agreement with the measured data.
(303) The results were a target rotation by dφ=+5.00°, with zoom on lower right display section of GUI showing dx=0.07 mm, dy=1.78 mm, dz=−0.01 mm, dφ=4.97°, dθ=−0.03°, and dψ=0.08°.
(304) Pitch Rotation
(305) The top rotation stage in the 6-DOF displacement assembly controls pitch rotation (patient's “nodding” direction). The pitch goniometer control knob is turned by +5° degrees, and the absolute position change is calculated from the resulting motion of the centroid camera coordinates. The calculated pitch is 4.95°, with less than 0.1° in yaw. The center of rotation for the Thorlabs GNL10 goniometer stage is 25.4 mm above the mount surface, while the target is only 6.4 mm above the stage. For the resulting −19 mm lever arm, the x-displacement due to a 5° rotation is −19 mm* sin(5°)=−1.66 mm, the y-displacement is 0.00 mm, and the z-displacement is −19 mm* [1−cos(5°)]=0.07 mm. These displacements are all within 0.1 mm of measured data.
(306) The results were a target pitch rotation by dθ=+5.00°, with zoom on lower right display section of GUI showing dx=−1.63 mm, dy=0.09 mm, dz=0.17 mm, dφ=0.21°, dθ=4.95°, and dψ=−0.07°.
(307) Laboratory Testing of an Embodiment
(308)
(309) Additional Marker Embodiments
(310) Various embodiments of optical trackers can be used with motion tracking systems disclosed herein. For example, the marker 110 illustrated in
(311) In some embodiments of markers, sub-targets comprise dark concentric sub targets on a high contrast diffusely reflective background. In some embodiments, the sub-targets comprise high contrast diffusely reflective concentric sub-targets on a dark background. In some embodiments, the concentric sub-targets comprise dark concentric sub-targets on a high contrast retro-reflective background. In some embodiments, the concentric sub-targets comprise high contrast retro-reflective concentric sub-targets on a dark background. Retro reflective material or surfaces are configured to reflect light back to a source with minimal scattering. Retro-reflective materials can provide benefits to a motion tracking system to, for example, in combination with the concentric sub targets, enable even faster and efficient processing. For example, a background or sub target with retro-reflective properties can provide additional information to the motion tracking system when the motion tracking system is analyzing an image of the target.
(312) Various embodiments as disclosed herein have been disclosed with reference to a plurality of sub targets, in some embodiments three sub targets, on a single substrate that is connected to the patient being monitored. Further, in some embodiments, such as the embodiments illustrated in
(313)
(314)
(315)
(316)
(317)
(318)
(319) In various embodiments wherein sub targets 2902 are each positioned on their own substrate 2904, the substrates 2904 can be permanently attached to a backing 3006, removably attached to a backing 2006, or setup in various other ways to be permanently attached together or removably attached. For example, a substrate 2904 may be connected to a backing 3006 using glue, other adhesive, hook and loop fasteners, screws, tape, and/or the like.
(320) Additional Process Flow Embodiments
(321)
(322) The process flow begins at block 3102. At block 3102, the system initializes a video source object. At block 3104, the system grabs a frame from the video source, such as a camera. At block 3106, the system analyzes the frame to locate target centroids, such as centroids of the sub targets of a marker. The target centroid locating sub process is shown with reference to blocks 3108 through 3132. At block 3108, the process flow varies depending on whether this is the first iteration of locating the target or whether it is a subsequent iteration. If it is the first iteration, the process flow proceeds to block 3110. At block 3110, the frame is converted to a binary image with a 50% threshold (e.g., any pixels with a value lower than 50% of maximum are set to zero and any pixels with a value greater than 50% of maximum are set to one). At block 3112, the system identifies in the binary image the largest white region and selects that region. At block 3114, the system sets all other regions in the binary image to zero. At block 3116, the system inverts the image.
(323) At block 3118, the system again converts to a binary image with a 50% threshold. At block 3120, the system identifies the second, third, and fourth largest regions. At block 3122, the system sets all other regions to zero. At block 3124, the system calculates centroids of the connected regions. At block 3126, the system sorts and averages centroid coordinates. The centroid coordinates are then passed on to the operation at block 3134.
(324) Returning to block 3108, if the process being performed at block 3106 is being performed subsequent to the first iteration, the process flow proceeds from block 3108 to block 3128. At block 3128, the system extracts boxes around the last known centroid coordinates. At block 3130, the system performs a weighted average to find a centroid in each box. At block 3132, the system adds a centroid offset to each of the previous coordinates to produce the final centroid coordinates. These coordinates are then passed on to the process at block 3134.
(325) At block 3134, the system calculates the 3-D position and rotation angles of the optical target. In some embodiments, the 3-D position and rotation angles are also converted from the optical detector coordinate system to, for example, an MRI machine coordinate system. At block 3136, the target position and rotation angles are passed onto an external process to enable motion compensation.
(326)
(327) The process flow illustrated in
(328) At block 3206, the system is configured to identify nonzero pixels on the bounding box perimeter. By identifying nonzero picture of pixels on the bounding box perimeter, the system can estimate where each of the sub targets is. As can be seen in block 3304 of
(329) At block 3210, the system is configured to extract region of interest (ROI) boxes each containing one sub target. In some embodiments, the region of interest box is size scaled by the size of the target bounding box. As can be seen in block 3306 of
(330) At block 3212, the system is configured to set pixel values below 30% of the pixel maximum value to zero. For example, as seen in block 3308 of
(331) Block 3310 of
(332) In some embodiments, a motion tracking system can be configured to run only one iteration of the centroid determination process illustrated in
(333) Additional Marker Embodiments
(334) Various embodiments of optical trackers or markers used with motion tracking systems have been described in the present disclosure. For example, some embodiments, such as the marker 110 illustrated in
(335) However, in some embodiments, it may be advantageous to use a marker that has no rotational ambiguity in order to allow for unique identification in the location and orientation of the subtargets in the marker. A marker with three identically-sized subtargets, with their centroids arranged as the vertices of an equilateral triangle, may be limited by rotational degeneracy. Such a marker may not have a unique pattern or arrangement for every in-plane rotation angle (e.g., rotations about an axis oriented normal to the plane of
(336)
(337) As mentioned above, in some embodiments, rotational asymmetry is achieved in the marker through an offset of one concentric target from the center of the marker along the vertical axis. This rotational asymmetry may enable tracking of the marker through 360-degree in-plane rotation with no degeneracy (e.g., no ambiguity as to the current in-plane rotation orientation throughout a full 360 degrees). In some embodiments, the shapes of the subtargets of the marker are elongated in the horizontal dimension to increase the horizontal cross section of the marker as it rotates away from a given camera. In some embodiments, the markers may comprise three or more separate subtargets arranged in a rotationally asymmetric pattern, with each subtarget comprising at least two concentric shapes. Additional shapes may be incorporated into the marker pattern for the purposes of unique marker identification.
(338) Since all five subtargets of the marker 3400 of
(339) Since large concentric pair 3404 is larger than the small concentric pairs 3402, the system may be able to more quickly identify large concentric pair 3402 than the other concentric pairs. The position of the large concentric pair 3404 may then be used to quickly locate the remaining concentric pairs—the four small concentric pairs 3402—through a variety of methods. One method for locating the remaining concentric pairs may be taking the four closest concentric pairs that surround the large concentric pair 3402, and then grouping all five concentric pairs as a marker. Thus, the large concentric pair 3402 may be used to identify the location of the marker quickly. Otherwise, the centroids of each concentric pair can be located but the specific subtarget each concentric pair corresponds to is unknown. The concentric pairs may in some embodiments be put into local groups to determine the subtarget for each concentric pair, such as through an algorithm like perimeter edge detection, described previously herein. This may in some embodiments add undesired computational time and/or complexity to the algorithm for identifying the location of the marker or uniquely identifying subtargets. Further details on marker pattern centroid detection are provided below in the discussion associated with
(340)
(341)
(342)
(343) There may be conflicting requirements associated with the size and design of the actual marker used. Making the marker as small as possible may be beneficial for the patient's convenience, and/or to enable more markers to be used. However, the marker may need to be large enough that the system may be able to accurately determine the centroids of the marker from a pixelated image taken of the marker. Furthermore, there may be an ideal spacing associated with the subtargets in a marker. Increased distances between subtargets in a marker may result in improved tracking, since the tracking algorithm observes differences associated with the movement of the centroids of the subtargets. The larger the marker is, the larger this distance is for a given motion, and thus the better the tracking will be. Further discussion on the size of a marker is provided above in the present disclosure. In some embodiments, it can be desirable to utilize a marker comprising outer subtargets having centroid locations that define a bounding square of a 14×14 millimeter size. In other embodiments, the bounding square (or rectangle) may comprise one or more sides having a length of 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, or more millimeters.
(344) Marker Pattern Centroid Detection
(345) Since the motion tracking system tracks the relative positions of markers, it may be useful for the system to use a process that can reliably identify the centroids of the subtargets (also referred to as reference points or concentric rings, herein) in a marker. It may also be useful to quickly and uniquely identify the subtargets that comprise a market. Various processes or methods may be used to identify the subtargets and the centroids of the subtargets. The system may use different processes or methods based on the marker patterns. Some marker designs may allow the system to use more efficient processes or methods when identifying the subtargets and the centroids of the subtargets.
(346)
(347) At block 3702, the system may binarize each marker image received. The system may set all pixels above a certain threshold to be white, while setting all pixels below that threshold to be black. The system would then proceed to block 3704 to detect all connected regions within each image. The connected regions may be defined as all groups of white pixels that are touching each other. However, in some embodiments, the system may be configured to detect connected regions of all groups of black pixels, such as when the marker comprises black subtargets on a white background.
(348) At block 3706, the system then filters out connected regions that are too small to be a marker component (such as those less than 4 pixels in some embodiments). At block 3708, the system may filter out connected regions that are too large to be a marker component (such as those greater than 2000 pixels in some embodiments). The system may perform block 3706 and block 3708 in parallel, or in some embodiments, the system may perform those two blocks sequentially. The system may perform block 3706 first and then block 3708 second, or the system may perform block 3708 first and then block 3706 second.
(349) At block 3710, the system may calculate the centroids of the remaining connected regions. At block 3712, the system may then compare the centroids of each connected region. Those regions with centroids within 1 pixel of each other are identified as concentric subtarget pairs at block 3714. In other embodiments, other thresholds may be used, such 0, 2, 3, 4, 5, or more pixels. At block 3716, the system may then average the centroids of the regions in each concentric pair and record them as a valid marker subtarget in block 3716.
(350) Alternatively, the system may utilize a somewhat different process if the marker has a large concentric ring, such as the marker embodiment illustrated in
(351) At block 3718, the system may then determine the general in-plane rotation of each marker by calculating the angle of the line segment connecting the overall centroid of the group of subtargets comprising the marker to the central subtarget. With the rotation angle determined, a rotation transform is applied to the marker at block 3720 to remap it to an upright position.
(352) At block 3722, the system may then sort the subtargets comprising the marker by their relative position to the central marker. At block 3724 a system may count all of the connected regions within the bounding box of each marker to uniquely identify each observed marker.
(353)
(354) Image 3802 is a raw greyscale image of a marker with rotational asymmetry, such as the marker in
(355) The system then binarizes image 3802 in order to obtain image 3804 for locating all white, connected regions (e.g., block 3702 of
(356) The system may then be left with image 3806, which shows all the connected regions that have not been discarded, as they are within the defined size range that the system is looking for. At this point, the system may then calculate the centroids of all remaining objects in the image (e.g., block 3710 of
Generalized Six-Degree-of-Freedom Tracking System
(357) Overview
(358) This disclosure previously described some embodiments in which a MRI Head Tracker (or other motion tracking system) may take real-time input from two 2D imaging sensors and analyze the data to determine and report motions in six degrees of freedom with minimal latency. Some of those embodiments are configured to detect and measure three centroid positions on the marker and utilize those positions to determine the position of the patient's head, by combining data from the two images of the marker to generate six unique principal quantities. One example of such a process is illustrated in
(359) In some embodiments, a generalized six-degree-of-freedom algorithm is utilized that enables any non-zero number of cameras (and/or any non-zero number of markers) to be used with the system. Such a six-degree-of-freedom tracking system may utilize 1 to n cameras with no particular restrictions on camera locations, except that the field of view of each camera covers some or the entire region of interest and that the six-degree-of-freedom position of each camera is known in a common arbitrary coordinate system. In some embodiments, an increase in the number of cameras and/or markers increases the overall tracking accuracy and/or redundancy. However, an increase in the number of cameras and/or markers may also cause an increase in processing power and/or processing time required to estimate a patient's pose. Accordingly, it can be desirable to optimize the number of cameras and/or number of markers used based on the available processing power, the available space in, on, or around the medical scanner for positioning therein of cameras, and/or the available space and/or mounting positions for markers on the patient being scanned. In some embodiments, such as in an MRI scanner where a patient is scanned within a generally cylindrical bore, the system can be configured to position a plurality of cameras about the bore, with the cameras positioned such that at least one camera (or any other number of cameras) can always see at least one marker (or any other number of markers) in any anticipated position of the patient during the scan.
(360) In some embodiments, the system may use data from every camera that can currently see a marker (which may or may not include every camera in the system). For example, if cameras 1, 3, and 5 can see a marker, then the system may use data from just those three cameras while ignoring data from other cameras. The system may automatically determine the position of the marker based on which cameras can see the marker. In some embodiments, the system may be able to seamlessly deal with blocking or obstructing of the viewpath of one or more cameras. For example, if there are initially three cameras that can see a marker but one camera's viewpath becomes blocked, such as by the patient's arm, then the system may be able to use the other two cameras with unobstructed viewpaths to determine the marker (or markers) position. In some embodiments, more than one marker may be tracked by some or all of the cameras or detectors.
(361) In some embodiments, the accuracy of the motion tracking system may be lower when one camera is used as compared to when more cameras are used, although the accuracy may still be sufficient to reduce or eliminate motion artifacts in the medical imaging scans. The accuracy may improve with two or more cameras, and the tracking accuracy may generally increase with additional numbers of cameras that can detect a marker (the same marker and/or other markers) at a given time. The generalized six-degree-of-freedom algorithm may allow the system to switch dynamically from camera to camera, or add and remove cameras from use.
(362) In some embodiments, the system utilizes a generalized set of principal quantities (e.g., the principal quantities may be used with various marker designs). The principal quantities may comprise in some embodiments the sums and differences in the X and the Y coordinates between every pair of targets from every image. For example, a marker may be imaged by both camera 1 and camera 2. The principal quantities would be the sum and difference in the X coordinate and the Y coordinate between every possible combination of pairs of targets available for a given marker. The system may construct a set of principle quantities for any number of markers based on the data available. For example, if there are six cameras being utilized by the system and three of the cameras can see the centroids of that marker, then the system may generate a full set of principal quantities based on all of that combined data. In this manner, the system may adjust the principal quantities based on new data available, such as if a fourth camera was able to see the centroids in the marker (or if only one camera were able to see a marker). Additional details regarding principal quantity calculations are given below.
(363) In some embodiments, the system may utilize six cameras. In some embodiments, the system may utilize four cameras located on the circumference of a circle on or about the body coil of an MR scanner, pointed inward with positions adjusted for best coverage of the region(s) of interest. The number of cameras used in the system may be chosen based on available processing power, space, optimum view paths, and/or the like. Additional cameras may improve accuracy and allow for redundancy, although in some embodiments additional cameras may result in additional processing power being needed for a desired processing time.
(364) In some embodiments, the system may need to know the exact position and orientation of all the cameras in the system (or at least a position and orientation within sufficient tolerances). This information may be obtained in some embodiments as part of a separate calibration process, an example of which is further discussed below. The system may also need to know the relative positions of all the targets or subtargets within a marker. The relative positions of all the targets within a marker may in some embodiments be known before-hand based on the design of the marker.
(365) Target Displacement Equations
(366) In some embodiments, the marker nominal centroid positions are determined from the geometrical characteristics of the marker. The full six-degree-of-freedom translation is composed of a 3-D displacement as well as a 3-axis rotation. The rotations are separated into roll, pitch and yaw (although other conventions may be used), and the translated position through rotation follows the Euler rotation matrix formulation as follows (using right-handed angle conventions). The x, y, and z displacement coordinates then follow the obvious independent translations:
(367)
(368) Decomposing the six independent translations from the absolute and relative displacements of the measured marker centroids is the subject of this effort. The 2D inverse problem is somewhat more difficult than the 3D problem, in that after the marker centroid projections in focal plane row and column are determined, significant degeneracies remain in the unfolding matrices for each camera. Combining the data from multiple cameras removes these degeneracies through a series of interrelated, nonlinear equations. The fastest procedure for solving this inverse problem is obtained by a variant of the Newton-Raphson method, whereby an approximate first-order solution is proposed and tested against the known (measured) centroid locations on the camera focal planes. The residual error is divided by the local derivatives with respect to each component of rotation and translation, to determine an iterative correction. The first-order solution is chosen by considering the features of the projected target pattern which are most strongly affected by a single rotation angle or displacement, and linearizing the inversion problem along these feature axes.
Six-Degree-of-Freedom Extraction Algorithm
(369) Image Characterization and Calculation of Principal Quantities
(370) In this embodiment, the markers each comprise 3 or more targets (plus additional identifying marks in some embodiments) and may be observed by 0 to n cameras, where n is the total number of active cameras in the system. In the case of a marker being obscured or moved outside of the field of view of all cameras, its position is unknown and not considered in any subsequent calculations until such time as the marker moves to a position where is can again be detected. In other cases, characterization of the images is achieved, in this embodiment, by calculating both the sum and difference of the column (X) and row (Y) coordinates between each target image and every other target image across all cameras. For a given marker, with targets, T and cameras, C, the set of principal quantities (PQ) characterizing each marker is given by:
PQ.sub.x={X.sub.T.sub.k,l ε[1:n]
C.sub.kC.sub.l contain valid marker images}
PQ.sub.y={Y.sub.T.sub.k,lε[1:n]
C.sub.kC.sub.l contain valid marker images}
PQ=PQ.sub.x∪PQ.sub.y
(371) In this manner, a set of principal quantities, PQ, may be constructed for each marker based on all available image data.
(372) Characterizing Global Variation in Principal Quantities with Six-Degree-of-Freedom Motions
(373) Partial derivatives relative to subject displacements and rotations (φ, θ, ψ, Δx, Δy, Δz), of the principal quantities described above, about the initial (non-displaced) position, are computed numerically. Here: Roll φ is right-handed rotation about the x-axis Pitch θ is right-handed rotation about the y-axis Yaw ψ is right-handed rotation about the z-axis Δx is toe-to-head direction Δy is left-to-right direction Δz is down-to-up direction
(374) Starting from an initial marker position in 3-D world space, defined as (φ, θ, ψ, Δx, Δy, Δz)=(0, 0, 0, 0, 0, 0), the initial world coordinates (x.sub.0i,y.sub.0i,z.sub.0i) are determined for each target, i, comprising the marker based on the known geometric configuration of the marker and definition of a convenient coordinate origin.
(375) Local partial derivatives of each of the principal quantities, with respect to each of the 6 degrees of freedom (roll, pitch, yaw, dx, dy, dz), are performed numerically by evaluating changes in these quantities for small increments in each degree of freedom. Changes in the target positions for specified motions along the six degrees of freedom are computed using the Euler rotation matrices and translation vector:
(376)
(377) Subsequently, the normalized pinhole camera model projections of the new marker target positions are determined using a geometric projection calculation. Given accurate knowledge of camera positions, the normalized projections are given by:
x.sub.i,j= tan α.sub.i,j sin β.sub.i,j,y.sub.i,j=±tan α.sub.i,j cos β.sub.i,j [2A]
(378) Here x and y.sub.i,j are the horizontal and vertical pinhole camera projections for translated target i projected onto the camera j sensor, where the angles α.sub.i,j and β.sub.i,j are the polar and azimuth angles locating target i, relative to the camera j focal axis. In order to maintain the convention of a counterclockwise sense of rotation for the azimuth angle for each of the camera boresight axes, equation 2 is divided into two mirrored functions: one for cameras to the patient's left side and the second for cameras to the patient's right side. The polar and azimuth angles are calculated from the target position world coordinates as follows:
(379)
(380) Here the point (x.sub.⊥,j,y.sub.⊥,j,z.sub.⊥,j) is the point of intersection between the camera optical axis and the plane perpendicular to the optical axis which includes the translated target position (x.sub.i,y.sub.i,z.sub.i):
x.sub.⊥,j=x.sub.0+κ(x.sub.cj−x.sub.0);y.sub.⊥,j=y.sub.0+κ(y.sub.cj−y.sub.0);z.sub.⊥,j=z.sub.0+κ(z.sub.cj−z.sub.0), [5A]
(381) with (x.sub.cj,y.sub.cj,z.sub.cj) defining the 3-D position of camera j, (x.sub.0,y.sub.0,z.sub.0) defining the nominal boresight position of all cameras at the un-displaced target center and the constant x based on geometric projection and given by:
(382)
(383) In equation [4], the inverse cosine function is taken to range from 0 to π, and the appropriate sign for β.sub.i,j is given by:
[(z.sub.cj−z.sub.0){(x.sub.cj−x.sub.0)(x.sub.⊥,j−x.sub.i)+(y.sub.cj−y.sub.0)(y.sub.⊥,j−y.sub.i)}−(z.sub.⊥,j−z.sub.i){(x.sub.cj−x.sub.0).sup.2+(y.sub.cj−y.sub.0).sup.2}]
(384) To numerically evaluate the partial derivatives of the principal quantities about the initialized target position, the un-displaced 3-D target position coordinates (x.sub.0i,y.sub.0i,z.sub.0i) are first projected to camera coordinates using equations [2A] through [6A] above, and initial values are computed for each of the principal quantities described above (most should be zero or near-zero at the starting position). Then small increments of roll, pitch, yaw, x-, y- and z-axis displacements are introduced one at a time; for each increment the new world coordinates and the new camera projections of the target positions are computed and the principal quantities are re-calculated. The change in each principal quantity is divided by the small angular or displacement increment to determine the partial derivative.
(385) Radial and tangential lens distortion effects are applied to the normalized projections according to the following equation:
(386)
(387) Where r.sup.2=x.sup.2+y.sup.2, and the radial distortion coefficients, a, b, c and the tangential distortion coefficients d, e, are determined by a separate lens calibration procedure. Lastly, the distorted projection coordinates are mapped to pixel values for each camera as follows:
(388)
(389) Where f.l. and s.sub.pix are the lens focal length and camera pixel pitch, and X.sub.0 and Y.sub.0 are the pixel coordinates of the center of the pixel array (i.e. for a 1280×1024 array, X.sub.0=640.5 and Y.sub.0=512.5, assuming the center of the upper left corner pixel is defined as having coordinates (1,1).).
(390) For instance, to determine the partial derivatives with respect to roll, the displacement vector (φ, θ, ψ, Δx, Δy, Δz)=(ôφ, 0, 0, 0, 0, 0) is introduced to the general displacement equation [1] to determine the translated target positions (x.sub.i,y.sub.i,z.sub.i): The conversion to camera coordinates (X.sub.i,j, Y.sub.i,j) is then performed using equations [2] through [6], and the principal quantities are calculated as outlined above. The difference between each principal quantity and the corresponding value of that quantity for the un-displaced calculation is divided by the small increment in roll, to give the partial derivative of each quantity with respect to roll. To determine partial derivatives with respect to pitch, the displacement vector (φ, θ, ψ, Δx, Δy, Δz)=(0, ôθ, 0, 0, 0, 0) is used to initiate the calculations, and so on for all six degrees of freedom.
(391) Each of these six repetitions produces one column of the global partial derivative matrix:
(392)
Determining First-Order Displacement Vector
(393) A first-order approximation to the displacement matrix is determined by multiplying the matrix of measured principal quantities, as determined above, by the inverse of the partial derivative matrix computed above:
(394)
Characterizing Local Variation in Principal Quantities with Six-Degree-of-Freedom Motions
(395) First order values for (φ, θ, ψ, Δx, Δy, Δz) determined above are entered into the translation equation [1A] to determine the corresponding translated 3-D target position (x.sub.i,y.sub.i,z.sub.i) for each of the target positions. These world coordinates are projected to camera coordinates (X.sub.i,j, Y.sub.i,j) using equations [2A] through [6A], and the principal quantities are re-calculated. These quantities are compared against the measured values of these quantities determined above, to create a residual error matrix:
(σ.sub.PQ.sub.
(396) Local partial derivatives of the principal quantities are calculated by introducing small increments in roll, pitch, yaw, x-, y- and z-axis displacements one at a time as before, but this time the increments are relative to the first-order displacement vector. For each increment, the new world coordinates and the new camera projections of the target positions are re-computed and the principal quantities are re-calculated. The change in each principal quantity is divided by the small angular or displacement increment to determine a local partial derivative. For instance, to calculate partial derivatives with respect to roll, the first-order displacement vector {φ.sub.0, θ.sub.0, ψ.sub.0, (Δx).sub.0, (Δy).sub.0, (Δz).sub.0} is replaced by {φ.sub.0+ôφ, θ.sub.0, ψ.sub.0, (Δx).sub.0, (Δy).sub.0, (Δz).sub.0} and resulting changes to each of the principal quantities is divided by to determine the local derivative with respect to roll. This is repeated for each of the six degrees of freedom.
(397) Each of these six repetitions produces one column of the new local partial derivative matrix:
(398)
Determining Coarse Correction to the First-Order Displacement Vector
(399) A coarse correction is computed to improve the first-order displacement vector and reduce residual error, by multiplying the residual error matrix determined above by the inverse of the local partial derivative matrix, also determined above:
(400)
(401) The first-order displacement vector is incremented by the coarse correction matrix to create a better approximation to the displacement vector: {φ.sub.0+Δφ,θ.sub.0+Δθ,ψ.sub.0+Δψ,(Δx).sub.0+Δ(Δx),(Δy).sub.0+Δ(Δy),(Δz).sub.0+Δ(Δz)}.
(402) Iterative Fine Correction to Determine Final Six-Degree-of-Freedom Displacement Vector
(403) Characterizing Local Variation in Principal Quantities with Six-Degree-of-Freedom Motions and Determining Coarse Correction to the First-Order Displacement Vector are repeated, starting with the coarse-corrected displacement vector, to determine a fine correction to the displacement vector. After this iteration, the resultant fine correction increments are added to the coarse-corrected vector to create the new 6-DOF displacement vector. These processes are repeated iteratively until the magnitude of each of the fine corrections is less than the desired threshold (typically 10.sup.−6 mm or degrees) or the number of iterations reaches a predetermined maximum (typically 10 iterations).
(404) One of skill in the art will recognize that the above-described six degree of freedom tracking algorithm is merely one way of implementing the tracking techniques disclosed herein, and various modifications may be made to this process without departing from the techniques disclosed herein.
(405) Multi-Marker Tracking
(406) Some embodiments of the motion tracking systems disclosed herein may simultaneously use multiple markers, such as to increase accuracy and/or redundancy. It may be useful for the system to be able to distinguish each marker from one another. One way of physically distinguishing markers from one another is described above, in which rotationally asymmetrical markers have varying amounts of non-concentric dots or other indicators on them. Although these markers are physically distinct, it may be desirable for the system to distinguish and uniquely identify the markers algorithmically. This algorithm may be generalized, such that the system may simultaneously handle any number of markers, including a greater amount of markers than the arrangements shown in
(407) In addition to accuracy and redundancy benefits, the tracking of multiple markers may be helpful in determining the presence of non-rigid facial movement. A marker may undergo rigid motion and move with the skeletal structure of the patient, or the marker may undergo non-rigid motion that is not indicative of structural motion of the patient's head or other feature being tracked (e.g., from relative movement of a patient's skin, cartilage, and/or the like that is not rigidly connected to the patient's skeletal structure). It may be helpful to distinguish what kind of motion a marker is undergoing, because non-rigid motion may lead to an erroneous correction by the motion tracking system if it is detected as rigid motion.
(408) Some embodiments of systems disclosed herein may keep track of a history of the relative positions of all the markers in the system. Once the six-degree-of-freedom coordinates for all visible markers are determined, coordinate transforms may be applied to remap the coordinate system origin sequentially to each of the markers, recording the resulting relative positions of the remaining markers in each case. In this manner, the six-degree-of-freedom relative position of each marker with respect to every other marker may be determined. A history of the relative position data over several seconds is maintained and may be used to calculate the means and standard deviations of relative positions of all markers. Each new set of relative position data may be compared to the historical mean and standard deviation, σ. If the new value is more than ±3σ from the historical mean (or another value in other embodiments), then it may be flagged as a non-rigid motion.
(409) With the use of only one marker, the marker may be subject to non-rigid motion but the system may in some embodiments be unable to determine it.
(410) With two markers, the system may be able to look at the relative positions in space of the two markers. The system may be able to compare the latest set of coordinates for the two markers against their history. If the two markers move more than 3 sigma than their historical relative position (or another value in other embodiments), then the system may determine that there is non-rigid motion. However, with two markers it may be difficult to tell if one of them (or both of them) are undergoing non-rigid motion, only that it is happening.
(411) With three or more visible markers, the system may be able to look at relative positions of each marker with respect to the other markers, and compare their history. The system may be able to determine that marker 2 is moving relative to other markers, but all the other markers are maintaining their relative positions to each other. This may mean that marker 2 is likely to be the non-rigid moving marker. The system may then focus on the other markers that are moving rigidly, and use data from those markers in order to enhance accuracy. As another example, if there are three visible markers and non-rigid motion flags are only raised in the relative positions between markers 1&2 and markers 2&3, then it is deduced that either marker 2 is moving non-rigidly or markers 1 &3 are moving non-rigidly with exactly the same motion (to within 3σ). Additional visible markers in the system would further reduce the already unlikely probability of the latter.
(412) The multi-marker tracking algorithm may also allow tracking to continue in situations when camera viewpaths are obscured, or objects move out of the field of view of the cameras. More specifically, if two or more markers are determined to be moving rigidly, then their position data may be combined to further improve tracking accuracy. This may be done by specifying one marker in the system as the ‘primary’ marker (this may be done automatically, for example, by assigning the lowest numbered marker in the system as the primary marker), whose six-degree-of-freedom position is reported, and mapping all of the other visible, rigidly moving markers to its position using the historical relative position data. Once mapped to the same coordinate position, the six-degree-of-freedom coordinates are averaged. This method has the additional benefit that if the primary marker becomes obscured or moves outside of the field of view of all cameras, its six-degree-of-freedom position is automatically reconstructed based on the visible markers. Thus, even if the primary marker happens to be obscured at that time, mapping other markers (moving rigidly and relative to each other) to the historical relative position of the primary marker allows for the primary marker's location to be reconstructed.
Tracking System Calibration
(413) Overview
(414) This disclosure previously described some embodiments utilizing an example handling of the lens calibration, which involved the correction of a barrel distortion in the lens based on prior lens calibration measurements. However, a more sophisticated lens calibration method may be used in combination with the generalized, six-degree-of-freedom extraction algorithm.
(415) In some embodiments, there may be a separate lens calibration that is performed with a checkerboard target (or in some embodiments another type of target). The checkerboard target may have a structure that comprises a set of nine marker structures arranged in a grid pattern (or a different number of marker structures). A user may reorient the checkerboard target based on where the desired zero origin should be, then the cameras may image the checkerboard target. The calibration algorithm may then be used to optimize camera positions instead of principal quantities.
(416) In some embodiments, a radial component and a tangential component of the barrel distortion may be stored in the camera, which may be specific to a particular camera-lens combination. The system may guess a location for the target in the world coordinate system, and then project the marker, based what may be expected to be seen in the images. The previously-calculated distortion may be applied, and then the projection may be converted into pixels and paired with the actual data that was collected from the sensor.
(417) Tracking System Calibration Setup
(418) After each lens is assembled to a camera and a lens calibration is performed to determine its focal length and distortion coefficients, the cameras are integrated with MR system and the integrated system is calibrated as described herein. Although this example is described with respect to cameras integrated into an MR system, similar calibration concepts may be utilized with other camera setups.
(419)
(420) The integrated camera positions (defined as the center of the focal planes) are taken to lie at (x.sub.cj,y.sub.cj,z.sub.cj). The cameras' (3902 and 3904) optical axes are pointed at the nominal target center position, (x.sub.0,y.sub.0,z.sub.0), along the vectors (x.sub.0−x.sub.cj)ê.sub.x+(y.sub.0−y.sub.cj)ê.sub.y+(z.sub.0−z.sub.cj)ê.sub.z. The focal planes then lie in the laboratory planes (x.sub.0−x.sub.cj)(x−x.sub.cj)+(y.sub.0−y.sub.cj)(y−y.sub.cj)+(z.sub.0−z.sub.cj)(z−z.sub.cj)=0.
(421) Now, by convention for definition of angles, the cameras are oriented (spun as necessary about their optical axes) such that the vertical (“Y”) axis of each focal plane lays in the laboratory plane z=z.sub.cj; these vertical axes are then defined by the lines
(x.sub.0−x.sub.cj)(x−x.sub.cj)+(y.sub.0−y.sub.cj)(y−y.sub.cj)=0, or equivalently x(x.sub.0−x.sub.cj)+y(y.sub.0−y.sub.cj)=x.sub.cj(x.sub.0−x.sub.cj)+y.sub.cj(y.sub.0−y.sub.cj).
The focal plane vertical “Y” axis unit vectors
(422)
are approximately antiparallel to the laboratory x axis, in the usual case where x.sub.0≈x.sub.cj. Taking the vector cross product of the optical axis unit vectors
(423)
with the camera vertical axis vectors gives the camera focal plane horizontal axis vectors:
(424)
In the simple embodiment for which x.sub.cj=x.sub.0, the “X” and “Y” axis unit vectors reduce to
(425)
(426) After the cameras are mounted in the system, a reference target is placed approximately at the desired camera system origin and roughly aligned to the desired camera coordinate system x and y axes.
(427)
(428) The camera coordinate system origin typically corresponds to a position at or near the MR system isocenter. The calibration target positioning and alignment can be easily performed using laser cross-hairs available in many MR systems. Camera parameters, aside from lens focal length and distortion coefficients, which have been measured previously, include physical location of focal plane centers and angular orientations relative to the reference target center: (x.sub.cj,y.sub.cj,z.sub.cj,α.sub.cj,γ.sub.cj,ω.sub.cj). Here α.sub.cj are the polar angles of the camera optical axes relative to nadir, γ.sub.cj are the azimuth angles of these optical axes, and ω.sub.cj define any further “twist” rotation of the cameras about their optical axes (relative to the nominal focal plane orientation convention described in the boxed text above).
(429)
(430)
(431) Nominal camera positions (x.sub.cj,y.sub.cj,z.sub.cj) are chosen with consideration for viewing access through various models of MRI head coils. By convention, the nominal polar angle is taken to be
(432)
the nominal azimuth angle is taken to be
(433)
(434) (using the usual convention for the arctangent function, for which the angle falls in the first quadrant, or 0 to 90 degrees, if both numerator and denominator are positive, in the second quadrant if the numerator is positive and denominator negative, and so on), and the nominal twist angle (relative to the reference orientation described in boxed text above) is taken to be
ω.sub.cj=0.
Tracking System Calibration Algorithm
(435) In some embodiments, a calibration algorithm starts by determining the point on the z=z.sub.c plane which projects to the center of the focal plane. This point is unaffected by camera twist, but depends on all other camera variables as follows:
(436)
(437) The point (x.sub.0,j, y.sub.0,j, z.sub.0) becomes the apparent coordinate origin for camera j, but the laboratory coordinates (x.sub.0,y.sub.0,z.sub.0) of the target center in its home position on the calibration station are fixed, by definition, at this absolute point in space. At the home position, or at any other setting for the six translation/rotation degrees of freedom, the location of the target centroids can be calculated precisely in laboratory coordinates, with the assumption of no positional error due to the confirmed accuracy of the calibration station.
(438) Defining several vectors which are used in determining the location on the camera focal planes (X.sub.t, Y.sub.t) of the image projection from an arbitrary reference target centroid point (x.sub.t,y.sub.t,z.sub.t) in laboratory coordinates:
(439) The vector from the apparent origin of coordinates for either camera to the camera itself is:
{right arrow over (R)}.sub.0,cj=(x.sub.cj−x.sub.0,cj)ê.sub.x+(y.sub.cj−y.sub.0,cj)ê.sub.y+(z.sub.cj−z.sub.0)ê.sub.z
The vector between the apparent origin and the true target position is:
{right arrow over (R)}.sub.t,cj=(x.sub.t−x.sub.0,cj)ê.sub.x+(y.sub.t−y.sub.0,cj)ê.sub.y+(z.sub.t−z.sub.0)ê.sub.z 4B
(440)
(441) The vector between the apparent origin and the point of intersection of the camera optical axis with the perpendicular plane that contains the target point (see figure above) is:
(442)
The unit vector running from the optical axis intercept to the target centroid in the perpendicular plane is
(443)
Simple trigonometry from the figure above shows that the polar angle between the camera's optical axis and the true target point is:
(444)
(445) The azimuth angle of the target point relative to the focal plane vertical axis is determined by projection. In any plane perpendicular to the optical axis, the focal plane principal axis ({circumflex over (X)} and Ŷ) unit vectors are given by equations (2) and (1) respectively; in the absence of camera twist, the focal plane azimuth angle β is determined through the dot product of the r.sub.⊥ vector with the projected vertical focal plane (Ŷ) vector. Adding the camera twist (minus sign by convention and definition only),
(446)
(447) The arc-cos function is assumed to be bounded at 0 and 180 degrees, and the appropriate sign multiplier for the azimuth angle is given by the sign of the projection (dot product) of the target perpendicular on the focal plane horizontal axis vector, as sign[β.sub.t,cj]=sign({circumflex over (r)}.sub.⊥,cj.Math.{circumflex over (X)});
sign[β.sub.t,cj]=sign[(z.sub.0j−z.sub.0){(x.sub.0,cj−x.sub.cj)(x.sub.t−x.sub.⊥,cj)+(y.sub.0,cj−y.sub.cj)(y.sub.t−y.sub.⊥,cj)}+(z.sub.t−z.sub.⊥,cj){(x.sub.0,cj−x.sub.cj).sup.2+(y.sub.0,cj−y.sub.cj).sup.2}] (left side)
sign[β.sub.t,cj]=−sign[(z.sub.cj−z.sub.0){(x.sub.0,cj−x.sub.cj)(x.sub.t−x.sub.⊥,cj)+(y.sub.0,cj−y.sub.cj)(y.sub.t−y.sub.⊥,cj)}+(z.sub.t−z.sub.⊥,cj){(x.sub.0,cj−x.sub.cj).sup.2+(y.sub.0,cj−y.sub.cj).sup.2}] (right side)
(448) The camera projections of these new target centroid positions are determined using a geometric projection calculation. The normalized pinhole camera projections are given by:
x.sub.t,cj= tan α.sub.t,cj sin β.sub.t,cj,y.sub.t,cj=±tan α.sub.t,cj sin β.sub.t,cj
(449) Radial and tangential lens distortion effects are then applied to the normalized projections according to the following equation:
(450)
(451) Where r.sup.2=x.sup.2+y.sup.2, and the radial distortion coefficients, a, b, c and the tangential distortion coefficients d, e, are determined by a separate lens calibration procedure. Lastly, the distorted projection coordinates are mapped to pixel values for each camera as follows:
(452)
(453) Where f.l. and are s.sub.pix the lens focal length and camera pixel pitch, and X.sub.0 and Y.sub.0 are the pixel coordinates of the center of the pixel array (i.e. for a 1280×1024 array, X.sub.0=640.5 and Y.sub.0=512.5, assuming the center of the upper left corner pixel is defined as having coordinates (1,1).).
(454) Here X.sub.0,cj and Y.sub.0,cj are the horizontal and vertical number of the central pixel column and row of camera j (assumed to intersect the optical axis of that camera), and f.l..sub.cj and s.sub.pix are the camera j lens focal length and pixel pitch.
(455) Tracking System Calibration Routine
(456) After assembly and rough system alignment, a calibration routine is run to determine residual errors in camera orientations (ôx.sub.cj,ôy.sub.cj,ôz.sub.cj,ôα.sub.cj,ôγ.sub.cj,ôω.sub.cj).
(457) A small error displacement ôx.sub.cj of a camera in the x-direction or a small error rotation ôγ.sub.cj of the camera optical axis about the laboratory “z” axis will result in the center of the reference target (i.e. home position) being imaged at a displaced position relative to the vertical center of the focal plane. Likewise, small error displacements ôy.sub.cj and/or ôz.sub.cj of camera position in the y- and/or z-directions will result in the center of the reference target being imaged at a position displaced relative to the horizontal center of the focal plane. By viewing multiple target features simultaneously, (e.g. the individual centroids of the reference target), image displacements corresponding to errors in camera position or orientation along six degrees of freedom can be distinguished using a similar method to that used for in the Head Tracker 6DOF extraction algorithm.
(458) The self-calibration algorithm considers the effects of small errors in each of the camera orientation parameters on the projected centroid positions seen by each camera. The reference target center defines the origin of the camera coordinate system and its orientation defines the coordinate system axes. This becomes the basis of measurement for camera position/orientation errors. The reference target itself is assumed to provide truth data relative to size scale. Thus, by definition, the true positions of the reference target centroids are known in laboratory coordinates, and by extension, the expected locations on the camera focal planes for projections of these points are also known.
(459) Comparison of the measured pixel positions (X.sub.t,cj,tY.sub.t,cj) of the each of the imaged centroids on each camera, with the expected projections gives 2N equations for the six unknown orientation errors (ôx.sub.cj,ôy.sub.cj,ôz.sub.cj,ôα.sub.cj,ôγ.sub.cj,ôω.sub.cj) for each camera, where N is the number of unique centroids in the reference target pattern. Like the run-time head tracker software, the calibration algorithm is based on computing the local derivatives of the (X.sub.t,cj,Y.sub.t,cj) centroid projections with respect to each error axis, and then performing a matrix inversion to determine the errors themselves. The calculation is performed iteratively until convergence to within an acceptable tolerance is achieved.
(460) The procedure outlined above only requires the end user to place and orient the reference target once and then start the calibration routine, after which all calculations are performed without additional user interaction. The resulting coordinates for all cameras are then stored in the system. Camera position calibrations need only be performed periodically or when changes to camera positions are suspected.
Additional Features
(461) Adapting to Addition/Removal of Cameras (e.g., Camera Viewpaths being Unblocked/Blocked)
(462) In some embodiments there may be a plurality of cameras. However, in certain situations one or more cameras in the plurality of cameras may stop working or have their viewpath obscured, or otherwise be incapable of detecting an image of a marker. It may be helpful for the system to utilize a generalized six-degree-of-freedom algorithm, as disclosed herein, that allows the system to continue performing when a camera goes down (e.g., by having its viewpath obscured). The same method or process may also enable the seamless addition of cameras (e.g., when a viewpath becomes unobstructed) into the motion compensation system. Thus, desirably the motion tracking is not disturbed when a camera goes down, and it is also not disturbed when the camera comes back online.
(463)
(464) Then, at block 4109 the system generates a set of principal quantities. The system calculates the set of principal quantities by calculating both the sum and difference of the X (column) and Y (row) coordinates between each target image and every other target image across all cameras. Thus, a set of principal quantities may be constructed for each marker based on all available image data.
(465) The system at block 4110 numerically evaluates the partial derivatives relative to subject displacements and rotations of the principal quantities. More specifically, local partial derivatives of each of the principal quantities, with respect to each of the six degrees of freedom, are determined numerically by evaluating changes in the principal quantities for small increments in each degree of freedom. The system at block 4112 may perform a calibration procedure, and that calibration may be used to help numerically evaluate the partial derivatives of the principal quantities about the initialized target position. In some embodiments, the calibration procedure is only performed once, or only periodically. The calibration procedure does not have to be performed every time a patient is tracked. The system obtains a partial derivative matrix at the end of block 4110.
(466) At block 4114, the system may then determine a first-order approximation of the displacement matrix by multiplying the matrix of measured principal quantities, obtained following block 4109, with the inverse of the partial derivative matrix obtained at the end of block 4110.
(467) At block 4116, the system may then generate a residual error matrix. The residual error matrix may be obtained from taking the first-order approximations, entering them into a translation equation to determine the corresponding translated 3-D target position for each of the target positions, obtaining 3-D world coordinates that are projected to camera coordinates, and re-calculating the principal quantities. These principal quantities may be compared against the measured values of the principal quantities which were obtained after block 4109, and the differences would reside in the residual error matrix. A local partial derivative matrix can then be obtained, with respect to each of the six degrees of freedom, by calculating the effect of small increments in each of the six degrees of freedom relative to the first-order displacement vector. For each increment, the new world coordinates and the new camera projections of the target positions are re-computed and the principal quantities are re-calculated. The change in each principal quantity is divided by the small angular or displacement increment to determine a local partial derivative. Repeating this for each of the six degrees of freedom yields the six-column local partial derivative matrix, as each repetition produces one column.
(468) At block 4118, the system may then compute a coarse correction to improve the first-order displacement vector and reduce residual error, by multiplying the residual error matrix from block 4116 with the inverse of the local partial derivative matrix from block 4116. The system may then increment the first-order displacement vector by the coarse correction matrix to create a better approximation of the displacement vector.
(469) At block 4120, the system may then perform an interative fine correction. The system continue looping back to block 4116 to determine a fine correction to the displacement vector. With each iteration, the resultant fine correction increments are added to the coarse-corrected vector to create the new displacement vector. The steps are repeated iteratively until the magnitude of each of the fine corrections is less than some threshold, or a number of pre-defined iterations have been performed. Afterwards, at block 4122 the system arrives at a final displacement vector, which may be associated with the position of the marker.
(470) Additional Process Flow
(471) If a patient is sitting still, or moving very slowly, noise in the tracking data may cause uncertainty or noise in the image correction, making it possible for the image correction to do more harm than good. It may be insufficient to simply turn off the image correction and keep the patient centered on the origin, since a patient may turn his head during the procedure and end up falling asleep in a position with a large offset. Thus, it may be helpful to perform an image correction that may be able to filter out any noise if there is very little motion in the patient.
(472) In some embodiments, prior to reporting the Six-Degree-of-Freedom marker position data (e.g., prior to sending tracking data to a controller of the scanner for compensation for motion), a noise filter or output filter may be applied to smooth the fluctuations in coordinate values when there is little to no active motion.
(473) In some embodiments, this filter comprises an application of an adaptive Kalman filter. The Kalman filter may use a constant velocity kinematic model, in which the process noise (that represents the uncertainty of the model) is in the form of an unknown acceleration. The magnitude of the process noise may control the relative weighting of the model vs. the measured data in the filter output.
(474) In some embodiments, the new, measured data may be compared to the average of the last 10 data points and a scale factor dependent on this difference may be applied to the process noise input of the Kalman filter. This may cause the model to be strongly preferred when the new data falls within 3σ of the recent average (of the last 10 data points), and continuously shifts preference toward the data as the difference increases beyond this range. Thus, if there is a large process noise, the filter will strongly weight the new data coming in and will hardly do any filtering at all of this data. If there is very little process noise, the filter may prefer the model over the data and perform a very strong smoothing or averaging of the data. Thus, the resulting filter strongly smooths the data when there is little to no motion by the patient, yet responds instantly to large changes in position without overshoot or lag. If the patient performs a large, instantaneous motion, the filter adapts right away and will go with the new data coming in that is associated with that large motion.
(475)
(476) Motion compensation system 4201 also utilizes one or more detectors 108. However, motion compensation 4201 is configured to use a six-degree-of-freedom algorithm that has been generalized so that motion compensation system 4201 may use any number, from 1 to n, of detectors 108. These detectors 108 may be configured to be used with any number of markers. The figure shows marker 1 (110-1), marker 2 (110-2), all the way up to marker n (110-n). These markers may have slight variations in them that allow the markers to be uniquely identifiable from one another. Each marker may at different times be imaged by zero, one, two, more than two, or all the detectors 108.
(477) Calibration engine 4206 may comprise a calibration database 4216, a detector calibration filter 4218, and a target calibration filter 4220. In some embodiments, calibration engine 4206 may be configured to calibrate the system at the initial startup of the system. In other embodiments, calibration engine 4206 may be used for at least some calibration procedures during some or all motion tracking procedures. The detector calibration filter 4218 may be configured to calibrate the detectors 108 to the motion compensation system. In some embodiments, detector calibration filter 4218 may be configured to allow a manufacturer or assembler of a motion compensation system to input information specific to each detector 108, such as values associated with barrel distortion for the detector 108. In some embodiments, target calibration filter 4220 may be configured to calibrate the system to one or more specific targets or markers, such as a grid of nine markers having the embodiment shown in
(478) Tracking engine 4234 may comprise a reference point filter 4236, a principal quantities generator 4238, a reference frame translator 4240, an error filter 4242, a convergence filter 4244, and an output filter 4246. The components of the tracking engine 4234 may be configured to operate similarly to the process flow described in association with
(479) For example, the reference point filter 4236 may be configured to analyze determine the locations of, and identify, the reference points in each marker. This information may be used to determine if any detectors 108 are not detecting or imaging a certain marker, and then exclude the image data from that detector from calculation. Principal quantitates generator 4238 may be configured to determine a set of principal quantities, which are obtained by calculating both the sum and difference of the X (column) and Y (row) coordinates between each target image and every other target image across all cameras. The principal quantities may be constructed for each marker based on all available image data.
(480) The reference frame translator 340 may be configured to convert between the two dimensional reference frame of each detector 108 and the three dimensional frame of the motion tracking system 4203. The error filter 4242 may be configured to analyze differences in principal quantities based on the visualized reference points and based on estimates of an object's pose to determine an amount of error between the two. The convergence filter 4244 may be configured to perform an iterative process to reduce an amount of error in an object pose estimate until an object pose estimate has an acceptable amount of error.
(481) Output filter 4246 may be the adaptive Kalman filter that was previously discussed. Output filter 4246 may be configured to detect when a patient is not moving or exhibiting little movement, and then perform a strong smoothing or averaging of the data in order to reduce the impact of noise. Output filter 4246 may also be configured to detect when a patient is making large movements, and respond the large movements without overshoot or lag. At block 4222, the coordinate system converter may remap the coordinate system to the same axes as the MR coordinate system, and convert to a quaternion representation of the angles.
Variations
(482) Specific embodiments have been described in detail above with emphasis on medical application and in particular MRI examination of a patient's head. However, the teachings of the present invention can be utilized for other MRI examinations of other body parts where movements of up to six degrees of freedom are possible. In addition medical procedures involving imaging devices other than MRI equipment (e.g., CT, PET, ultrasound, plain radiography, and others) may benefit from the teaching of the present invention. The teachings of the present invention may be useful in many non-medical applications where tracking of a target having several degrees of freedom are possible. Some of these applications could be military applications. Furthermore, while particular algorithms are disclosed, variations, combinations, and subcombinations are also possible.
(483) Computing System
(484) In some embodiments, the computer clients and/or servers described above take the form of a computing system 1500 illustrated in
(485) Motion Correction Control Systems
(486) In an embodiment, the system 700 comprises a motion correction control system module 1514 that carries out the functions described herein with reference to motion correction mechanism, including any one of the motion correction methods described above. The motion correction control system module 1514 may be executed on the computing system 1500 by a central processing unit 1504 discussed further below.
(487) In general, the word “module,” as used herein, refers to logic embodied in hardware or firmware, or to a collection of software instructions, possibly having entry and exit points, written in a programming language, such as, for example, COBOL, CICS, Java, Lua, C or C++ or Objective C. A software module may be compiled and linked into an executable program, installed in a dynamic link library, or may be written in an interpreted programming language such as, for example, BASIC, Perl, or Python. It will be appreciated that software modules may be callable from other modules or from themselves, and/or may be invoked in response to detected events or interrupts. Software instructions may be embedded in firmware, such as an EPROM. It will be further appreciated that hardware modules may be comprised of connected logic units, such as gates and flip-flops, and/or may be comprised of programmable units, such as programmable gate arrays or processors. The modules described herein are preferably implemented as software modules, but may be represented in hardware or firmware. Generally, the modules described herein refer to logical modules that may be combined with other modules or divided into sub-modules despite their physical organization or storage.
(488) Computing System Components
(489) In an embodiment, the computing system 1500 also comprises a workstation or other computing devices suitable for controlling and/or communicating with large databases, performing transaction processing, and generating reports from large databases. The computing system 1500 also comprises a central processing unit (“CPU”) 1504, which may comprise a conventional microprocessor. The computing system 1500 further comprises a memory 1508, such as random access memory (“RAM”) for temporary storage of information and/or a read only memory (“ROM”) for permanent storage of information, and a mass storage device 1502, such as a hard drive, diskette, or optical media storage device. Typically, the modules of the computing system 1500 are connected to the computer using a standards based bus system. In different embodiments, the standards based bus system could be Peripheral Component Interconnect (PCI), Microchannel, SCSI, Industrial Standard Architecture (ISA) and Extended ISA (EISA) architectures, for example.
(490) The computing system 1500 comprises one or more commonly available input/output (I/O) devices and interfaces 1512, such as a keyboard, mouse, touchpad, and printer. In one embodiment, the I/O devices and interfaces 1512 comprise one or more display devices, such as a monitor, that allows the visual presentation of data to a user. More particularly, a display device provides for the presentation of GUIs, application software data, and multimedia presentations, for example. In the embodiment of
(491) Computing System Device/Operating System
(492) The computing system 1500 may run on a variety of computing devices, such as, for example, a mobile device or a server or a desktop or a workstation, a Windows server, an Structure Query Language server, a Unix server, a personal computer, a mainframe computer, a laptop computer, a cell phone, a personal digital assistant, a kiosk, an audio player, a smartphone, a tablet computing device, and so forth. The computing system 1500 is generally controlled and coordinated by operating system software, such as iOS, z/OS, Windows 95, Windows 98, Windows NT, Windows 2000, Windows XP, Windows Vista, Windows 7, Linux, BSD, SunOS, Solaris, or other compatible operating systems. In Macintosh systems, the operating system may be any available operating system, such as MAC OS X. In other embodiments, the computing system 1500 may be controlled by a proprietary operating system. Conventional operating systems control and schedule computer processes for execution, perform memory management, provide file system, networking, and I/O services, and provide a user interface, such as a graphical user interface (“GUI”), among other things.
(493) Network
(494) In the embodiment of
(495) Access to the motion correction control system module 1514 of the computer system 1500 by computing systems 1520 and/or by data sources 1522 may be through a web-enabled user access point such as the computing systems' 1520 or data source's 1522 personal computer, cellular phone, laptop, or other device capable of connecting to the network 1518. Such a device may have a browser module is implemented as a module that uses text, graphics, audio, video, and other media to present data and to allow interaction with data via the network 1518.
(496) The browser module may be implemented as a combination of an all points addressable display such as a cathode-ray tube (CRT), a liquid crystal display (LCD), a plasma display, touch screen display or other types and/or combinations of displays. In addition, the browser module may be implemented to communicate with input devices 1512 and may also comprise software with the appropriate interfaces which allow a user to access data through the use of stylized screen elements such as, for example, menus, windows, dialog boxes, toolbars, and controls (for example, radio buttons, check boxes, sliding scales, and so forth). Furthermore, the browser module may communicate with a set of input and output devices to receive signals from the user.
(497) The input device(s) may comprise a keyboard, roller ball, pen and stylus, mouse, trackball, voice recognition system, or pre-designated switches or buttons. The output device(s) may comprise a speaker, a display screen, a printer, or a voice synthesizer. In addition a touch screen may act as a hybrid input/output device. In another embodiment, a user may interact with the system more directly such as through a system terminal connected to the score generator without communications over the Internet, a WAN, or LAN, or similar network.
(498) In some embodiments, the system 1500 may comprise a physical or logical connection established between a remote microprocessor and a mainframe host computer for the express purpose of uploading, downloading, or viewing interactive data and databases online in real time. The remote microprocessor may be operated by an entity operating the computer system 1500, including the client server systems or the main server system, an/or may be operated by one or more of the data sources 1522 and/or one or more of the computing systems. In some embodiments, terminal emulation software may be used on the microprocessor for participating in the micro-mainframe link.
(499) In some embodiments, computing systems 1520 that are internal to an entity operating the computer system 1500 may access the motion correction control system module 1514 internally as an application or process run by the CPU 1504.
(500) User Access Point
(501) In an embodiment, the computing system 1500 comprises a computing system, a smartphone, a tablet computing device, a mobile device, a personal computer, a laptop computer, a portable computing device, a server, a computer workstation, a local area network of individual computers, an interactive kiosk, a personal digital assistant, an interactive wireless communications device, a handheld computer, an embedded computing device, or the like.
(502) Other Systems
(503) In addition to the systems that are illustrated in
(504)
(505) The electronics package 1710 is configured to detect the location of the fiducial marker or other marker 1702 in order to determine the location of the diseased tissue 1704. In an embodiment, the electronics package 1710 is coupled to a first marker tracking system 1708. The first marker tracking system 1708 can be configured to receive tracking data from the electronics package 1710 to determine the location of the markers 1702. By determining the locations of the markers 1702, the first marker tracking system 1708 can be configured to determine the location of the diseased tissue 1704. In an embodiment, the systems and methods disclosed herein for tracking markers with 0.1 mm and 0.1 degree accuracies can be implemented or employed by the first marker tracking system 1708. In an embodiment, the electronics package 1710 comprises an optical marker 1712. The optical marker 1712 is configured to be detected by an optical scanner 1714, for example, a CCD camera. In an embodiment, the optical scanner 1714 is coupled to an optical marker tracking system 1716.
(506) The optical marker tracking system 1716 can be configured to determine the location of the electronics package 1710 relative to the therapeutic application system 1718. In an embodiment, the systems and methods disclosed herein for tracking markers with 0.1 mm and 0.1 degree accuracies can be implemented or employed by the optical marker tracking system 1716. In an embodiment, the system can comprise a coordinate generating system 1720 that is configured to receive tracking data from the first marker tracking system 1708 and the optical marker tracking system 1716. The coordinate generating system 1720 can be configured to analyze the tracking data in order to generate coordinate data that can be used to identify the location of the diseased tissue 1704. In an embodiment, the coordinate generating system 1720 can be configured to transmit the coordinate data to the therapeutic application system 1718.
(507) The therapeutic application system 1718 can be configured to generate a therapeutic beam based on the coordinate data. For example, the therapeutic application system 1718 can be configured to direct a radiation beam to a particular location in the patient 1706 based on the coordinate data. Further, the therapeutic application system 1718 can also be configured, for example, to generate a particular radiation beam shape based on the coordinate data. Any patient movement can be detected by the electronics package 1710 and the optical scanner 1714. The first marker tracking system 1708 and the optical marker tracking system 1716 can be configured to generate new tracking data to be inputted into the coordinate generating system 1720. The coordinate generating system 1720 can be configured to generate new coordinate data for transmission into the therapeutic application system 1718. The therapeutic application system 1718 can be configured to analyze the new coordinate data in order to redirect and/or reshape the therapeutic beam to be applied to the diseased tissue 1704 of the patient 1706.
(508) Similar to
(509)
(510) At block 1818, the system can be configured to track the optical marker position in order to determine the location of the electronics package. At block 1816, the system can be configured to determine the position of the electronics package relative to the therapeutic equipment. At block 1818, the system can be configured to analyze the tracking data of the first marker and the optical marker, and generate coordinates of the target tissue site. The coordinates of the target tissue site can be transmitted to the therapeutic therapy equipment at block 1820. The therapeutic therapy equipment can be configured to utilize the coordinate data to transmit the therapeutic therapy to the surgical site at block 1820. At decision block 1822, the system can be configured to repeat the application of the therapeutic therapy. If the therapeutic therapy application should be repeated, the system can be configured to loop back to block 1806 to initiate the tracking systems. If the therapeutic therapy application should not be repeated, the system can be configured to end the process at block 1824.
(511) Similar to
(512) Conditional language, such as, among others, “can,” “could,” “might,” or “may,” unless specifically stated otherwise, or otherwise understood within the context as used, is generally intended to convey that certain embodiments include, while other embodiments do not include, certain features, elements and/or steps. Thus, such conditional language is not generally intended to imply that features, elements and/or steps are in any way required for one or more embodiments or that one or more embodiments necessarily include logic for deciding, with or without user input or prompting, whether these features, elements and/or steps are included or are to be performed in any particular embodiment. The headings used herein are for the convenience of the reader only and are not meant to limit the scope of the inventions or claims.
(513) Although this invention has been disclosed in the context of certain preferred embodiments and examples, it will be understood by those skilled in the art that the present invention extends beyond the specifically disclosed embodiments to other alternative embodiments and/or uses of the invention and obvious modifications and equivalents thereof. Additionally, the skilled artisan will recognize that any of the above-described methods can be carried out using any appropriate apparatus. Further, the disclosure herein of any particular feature, aspect, method, property, characteristic, quality, attribute, element, or the like in connection with an embodiment can be used in all other embodiments set forth herein. For all of the embodiments described herein the steps of the methods need not be performed sequentially. Thus, it is intended that the scope of the present invention herein disclosed should not be limited by the particular disclosed embodiments described above.