Device For Imaging Blood Vessels

Abstract

A device for automatically imaging the capillary blood vessels of a living tissue likely to move, configured for selecting images of the sequence, called ‘sharp images’, arranged in chronological order of acquisition, shuffling the sharp images, for decorrelating temporally the sharp images, by arranging them in a shuffled order different from the chronological order, realigning spatially the sharp images arranged in the shuffled order, generating a projected image by projection of the pixels of the realigned sharp images, in a stack, the projected values of the pixels forming the projected image being extremal intensity values of the pixels of all the sharp images, the projection of the extremal of intensity values of the pixels rendering all the positions of all erythrocytes of all the sharp images in the projected image.

Claims

1. A device for automatically imaging the capillary blood vessels of a living tissue likely to move, comprising: a light source; an optical device to guide the light towards the living tissue, the light being absorbed by erythrocytes of the capillary blood vessels; a camera for acquiring at least one sequence of images which shows a region of interest of the capillary blood vessels, over a given duration, the absorption of the light by the erythrocytes on the images showing the capillary blood vessels, a microscope connected to the camera, a process and display unit, connected to the camera, the process and display unit being configured for: selecting images of the sequence, called ‘sharp images’, arranged in chronological order of acquisition; shuffling the sharp images, for decorrelating temporally the sharp images, by arranging them in a shuffled order different from the chronological order; realigning spatially the sharp images arranged in the shuffled order; generating a projected image by projection of the pixels of the realigned sharp images, in a stack, the projected values of the pixels forming the projected image being: minimal intensity values of the pixels of all the sharp images, the projection of the minimal of intensity values of the pixels rendering all the positions of all erythrocytes of all the sharp images in the projected image, or average intensity values of the pixels of all the sharp images, the projection of the average intensity values rendering the average flow in the capillary vessels in the projected image, or maximal intensity values of the pixels of all the sharp images, the projection of the maximal of intensity values of the pixels rendering the positions where the flow of erythrocytes is stagnant.

2. The device according to claim 1, wherein the process and display unit is configured for re-aligning spatially the sharp images in the shuffled order, without taking into account a fixed common vascular pattern of all the sharp images.

3. The device according to claim 1, wherein the process and display unit is configured to project at least some of the sharp images in their chronological order on the projected image to see a video of the flow of erythrocytes in the capillary blood vessels.

4. The device according to claim 1, wherein the process and display unit is configured to shuffle the sharp images such that two successive images in the shuffled order correspond to images which are separated one with respect to the other by at least two other images successive in the chronological order, preferably by more than 50 images in the chronological order, preferably corresponding to at least one second in actual time.

5. The device according to claim 1, wherein the process and display unit is configured to shuffle the sharp images by: separating the sequence of sharp images into at least two subsequences, distributing sharp images of the subsequences to define the sharp images of the shuffled order such that successive positions of the shuffled order correspond to sharp images from the different subsequences.

6. The device according to claim 1, wherein the device comprises a control unit coupled to the camera, the control unit being configured to control the camera to acquire a plurality of sequences, each sequence being acquired during a given time duration, and to control a given length of time between each sequence.

7. The device according to claim 6, wherein the process and display unit is configured to calculate the projected images from the respective sequences, arranged in the chronological order of acquisition of the sequences, one projected image per sequence, so as to display a video of the evolution of the capillary blood vessels with the projected images at least some of the sharp images being projected in their chronological order on the projected images to see the video of the flow of erythrocytes in the capillary blood vessels.

8. The device according to claim 6, wherein the control unit is configured to move the camera if the region of interest moves without returning to its initial position, so as to continue to observe the region of interest after its displacement with respect to the initial position in the living tissue.

9. Device according to claim 1, wherein, for obtaining the sharp images, the process and display unit is configured to: calculate the minimal reference gray level of the reference image; select the images of the sequence which have a minimal gray level close to the reference gray level by defining a criterion such that a given number of images, for example 50% are considered to be close to the reference image and conserved; calculate the total intensity by summing all the pixels of the first selected images; select the images among the first selected images which have a total intensity above a threshold, the second selected images selected being the sharp images.

10. The device according to claim 1, wherein the process and display unit is configured to: measure the resulting intensity of the all pixels in the capillary blood vessels on the projected image by the minimal values projection, and calculate the depth of the capillary blood vessels perpendicular to the planes of the images; the depth being a function of the intensity of the pixels (which is function of the number of erythrocytes in the Z-direction, which moves in the blood vessels) so as to deduce the cross section of the capillary blood vessels.

11. The device according to claim 1, wherein the spatial realignment comprises shifting spatially and/or rotating each sharp image with respect to the previous sharp image arranged in the shuffled order.

12. The device according to claim 1, wherein the camera comprises an image sensor and an optical filter which allows transmission to the image sensor of filtered light at wavelengths comprised between 450 nm and 650 nm, preferably between 490 nm and 590 nm.

13. The device according to claim 1, wherein the sequence comprises at least 30 sharp images and preferably at least 100 sharp images for the projected image, the camera being configured to acquire at least 10 images of the sequence per second.

14. Method for automatically imaging capillary blood vessels of a living tissue likely to move, comprising: illuminating the living tissue, the light being partially absorbed by the erythrocytes of the capillary blood vessels; acquiring at least one sequence of images which shows a region of interest of the capillary blood vessels, over a given duration, selecting images of the sequence, called ‘sharp images’, arranged in chronological order of acquisition; shuffling the sharp images, for decorrelating temporally the sharp images, by arranging them in a shuffled order different from the chronological order; realigning spatially the sharp images arranged in the shuffled order; generating a projected image by projection of the pixels of the realigned sharp images, in a stack, the projected values of the pixels forming the projected image being: minimal intensity values of the pixels of all the sharp images, the projection of the minimal of intensity values of the pixels rendering all the positions of all erythrocytes of all the sharp images in the projected image, or average intensity values of the pixels of all the sharp images, the projection of the average intensity values rendering the average flow in the capillary vessels in the projected image or maximal intensity values of the pixels of all the sharp images, the projection of the maximal of intensity values of the pixels rendering the positions where the flow is stagnant

15. The method according to claim 14, wherein the spatial re-alignment of the sharp images in the shuffled order, is realized without taking into account a fixed common vascular pattern of all the sharp images.

16. The method according to claim 14, wherein after the processing of the projected image, the method comprises the step of projecting at least some of the sharp images in their chronological order on the projected image to see a video of the flow of erythrocytes in the capillary blood vessels.

17. The method according to claim 15, wherein the shuffling comprises the two following sub-steps: separating the sequence of sharp images into at least two subsequences, distributing sharp images of the subsequences to define the sharp images of the shuffled order such that successive positions of the shuffled order correspond to sharp images from different subsequences.

18. The method according to claim 15, wherein the camera is controlled to acquire a plurality of sequences, each sequence being acquired during a given duration, and to wait for a given length of time between each sequence, the method comprises a step of calculating the projected images from the respective sequences, arranged in the chronological order of acquisition of the sequences, one projected image per sequence, so as to display a video of the evolution of the capillary blood vessels with the projected images.

19. The method according to claim 15, wherein the method comprises the steps of: measuring the intensity of the pixels in a capillary blood vessel on the projected image, and calculating the depth of the capillary blood vessels perpendicular to the planes of the images; the depth being a function of: the intensity of the pixels and the width of the capillary blood vessels, so as to deduce the cross section of the capillary blood vessel.

20. Use of a device according to claim 1 for in vitro/in vivo imaging capillary blood vessels of a living tissue, the living tissue being chosen among: the chick chorioallantoic membrane, mouse or rat's ear, the rabbit mesentery, the zebrafish vessels, chicken brain, chicken retina, chicken superficial vasculature of the limb, chicken yolk-sac, and generally, all embryos developing in ovo (reptiles, birds) and having similar extra-embryonic organs and visible blood vessels (e.g. limb vessels, intersomitic vessels etc.), the retina.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The invention will be described by way of example, with reference to the accompanying drawings in which:

[0075] FIG. 1 illustrates a device according to a possible embodiment of the invention,

[0076] FIG. 2 illustrates different steps of image processing,

[0077] FIG. 3, FIG. 4 and FIG. 5 are images of a CAM vasculature acquired and processed by the device,

[0078] FIG. 6 is an image of the terminal arteriole acquired and processed by the device,

[0079] FIG. 7 is an image of a CAM vasculature acquired and processed by the device presenting hotspots,

[0080] FIG. 8 is an image of a CAM vasculature acquired and processed by the device,

[0081] FIG. 9 and FIG. 10 illustrate terminal segments and proximal segments of arterioles,

[0082] FIG. 11, FIG. 12, FIG. 13 and FIG. 14 illustrate the optical profile of vessels,

[0083] FIG. 15 is an image of 3D collaterals acquired and processed by the device,

[0084] FIG. 16 is an image of the vascular interlace at day 10 acquired and processed by the device,

[0085] FIG. 17 and FIG. 18 are images of a CAM vasculature acquired and processed by the device presenting hotspots by selecting minimums or average values of pixels,

[0086] FIG. 19 is an image of the CAM vasculature acquired by the camera,

[0087] FIG. 20 is an image of the CAM vasculature acquired and processed by the device,

[0088] FIG. 21 illustrates a process for measuring the depth of a blood vessel,

[0089] FIG. 22 is an image of the CAM vasculature acquired and processed by the device.

DEFINITION

[0090] The term “realignment” of images will be used herein to designate the process of transforming different images with a relative motion with respect to each other into one coordinate system of images in which the relative movements are cancelled. The term “realignment” of images is also known as “registering” images.

DETAILED DESCRIPTION OF PREFERRED ASPECTS OF THE INVENTION

[0091] Referring to FIG. 1, the device 1 is configured for automatically imaging the capillary blood vessels and also larger vessels in the hierarchy of vessels, of a living tissue 2, for example an embryo.

[0092] Embryo

[0093] The living tissue 2 can be for example an embryo. Prior to imaging, the embryo is incubated shell less in a plastic cup 10. The plastic cup 10 has a trapezoidal profile, which is important to image the Chorioallantoic Membrane (CAM) vessels by the edge, over a white background. The egg was opened and transferred to the plastic cup in a sterile hood. The shell of the egg was sterilized before being broken. The cup was placed in a large Petri dish (Duroplan 10 cm), with optical quality. 3 mm of PBS were poured around the plastic cup inside the Petri dish. The Petri dish comprising the embryo was incubated in an incubator at 37° C. For intra-vital imaging, the embryos, inside the Petri dish, were placed in between two heating stages (Minitüb gmbh). The top heating stage has a central round window.

[0094] Device

[0095] The device 1 comprises a light source 3. The light source 3 can be any light source powerful enough (−1500 watts, e.g. fiber lamp Schott). The light is oriented towards the sample by, for example, a light fiber. The light is partially absorbed by the living tissue 2 and especially by the erythrocytes circulating in the capillary blood vessels of the living tissue 2. This will provide the contrast for the image processing.

[0096] The device 1 comprises a camera 5 for acquiring at least one sequence of images of a region of interest 6 (ROI) of the capillary blood vessels, over a given duration. The camera 5 can be a Stingray monochrome HD firewire camera from Allied Vision Technology (frame rate 15 Hz), or a Basler CMOS monochrome HD USB camera (frame rate 42 Hz), or any other camera providing computerized digital images.

[0097] The device 1 comprises a microscope 7 coupled to the camera 5. The microscope 7 can be a Leica Macroscope F16 APO. The device 1 also comprises a process and display unit 8 connected to the camera 5/microscope 7, so as the camera 5 can transfer at least an image sequence to the process and display unit 8. The sequence comprises images acquired by the camera 5 in a chronological order.

[0098] The microscope 7, the camera 5 and the living tissue 2 are arranged so that the camera 5 can image a region of interest 6 of the living tissue chosen by the user of the device 1.

[0099] Preferably, the device 1 comprises an optical filter which permits transmission to the image sensor of filtered light at wavelengths comprised between 450 nm and 650 nm, preferably between 490 nm and 590 nm. Therefore, the imaged erythrocytes correspond to a minimum of measured intensity.

[0100] Process and Display Unit 8

[0101] Referring to FIG. 2, the process and display unit 8 is configured for performing different steps of image processing.

[0102] The living tissue 2, when moving, can indeed be displaced in large proportions, or moved in a direction normal to the plane of observation. The image of the region of interest 6 acquired by the camera 5 can get blurred. Therefore, in a step 201, the process and display unit 8 selects sharp images from the image sequence acquired by the camera 5. Generally, blurred images arise from the oscillatory behavior of the heartbeat of the living tissue and intrinsic movements of the embryo. For example, with proper positioning of the objectives, the region of interest 6 can be acceptably focused for 50% of the time, and out-of-focus 50% of the time, due to the oscillation of the living tissue 2. Approximatively 1 out of 2 plates can then be discarded. When a sharp feature gets out of focus, the light is diffused away, so that typically a sharp dark spot will become lighter in color or in gray level. A crisp and dark area can be selected (automatically or by user) in a reference image of the sequence acquired by the camera 5. The process and display unit 8 is configured to measure the gray level of the area image-by-image and discard for example 50% of the images having a less sharp area, as deduced from measuring the gray level.

[0103] The grey level of a selected area can be measured, for example, by calculating the minimal gray level of the selected area. Then, the process and display unit 8 can select first selected images from the sequence, each image of the first selected images having a minimal gray level comprised between the minimal gray level of the reference image and a threshold grey level being higher than the minimal gray level of the reference image. The grey level of a selected area can also be measured by averaging the grey levels of the area. Then, the process and display unit 8 is configured to select the images presenting an average grey level comprised within a given range of grey levels, comprising the average grey level of the reference image.

[0104] The sharp images can also be selected by the process and control unit 8 among the images of the sequence by calculating the intensity gradient of each image. Therefore, only the image presenting an intensity gradient superior than a given threshold value can be selected.

[0105] The region of interest 6 of the living tissue 2, when moving, can also be displaced out of the field of view seen by the camera 5. Therefore, in a step 202, images which are too different from a chosen reference image in the sequence acquired by the camera 5. The difference between a reference image, chosen by the user, and an image of the sequence can be measured by the same means than in step 201. The process and display unit 8 can then discard some of the images of the sequence for a given threshold grey level.

[0106] In a step 203, the images from the output sequence of step 201 and/or of step 202 are shuffled for decorrelating temporally the images, by arranging them in a shuffled order different from the chronological order. The camera 5 can, for example, acquire a sequence comprising between 50 and 300 images (for example at 15 hz) which amounts to 5 to 20 seconds of video acquisition. After step 203, two successive images in the shuffled order correspond to images which are separated one with respect to the other by at least two other images successive in the chronological order, notably by more than 10 images and preferably by more than 25 images in the chronological order, preferably corresponding to at least one second in actual time.

[0107] The process and display unit 8 can be configured for separating the output sequence of steps 201 and/or of step 202 in two sub-sequences and distributing sharp images of the subsequences to define the sharp images of the shuffled order such that successive positions of the shuffled order correspond to sharp images from the different subsequences.

[0108] In order to do so, the process and display unit 8 can, for example, separate the output sequence of step 201 and/or 202 in a first subsequence and a second subsequence of equal number of images (in the case of an even number of images in the output sequence of step 201 and/or 202). For N.sub.im being the number of images in the sequence, k being an integer, the process and display unit 8 can be configurated for shuffling the positions of each 2k+1 image between 1 and N.sub.im/2 of the first subsequence respectively by each N.sub.im/2+k image of the second subsequence. Therefore, the temporal decorrelation between the images of the sequence can be maximized.

[0109] Alternatively, the images of the output sequence of step 201 and/or 202 can be randomly shuffled.

[0110] In a step 204, the images arranged in the shuffled order, i.e. the images of the output sequence of step 203, are realigned (i.e. are registrated). The output sequence of step 203 is for example aligned using the plugin StackReg (Biomedical Imaging Group, EPFL) in the software ImageJ. Each image is used as the template with respect to which the next image is aligned, so that the alignment proceeds by propagation. The user can select the reference image from which the sequence is aligned. Each image of the sequence can be aligned towards the previous image by translation, by rigid body transformation, by scaled rotation and/or by affine transformation.

[0111] The realignment of the step 204 is made possible by the step 203 of shuffling. The region of interest 6 presents two types of movements: the background movement of the living tissue 2 and the movement of the flowing erythrocytes in the blood vessels or capillaries. If realigning a sequence without shuffling the different images, the realignment process can consider the two types of movement between two consecutive images, and therefore align considering the movement of the erythrocytes. By shuffling the images of the sequence in step 203, the movement of the erythrocytes is not detectable between two consecutive images, and therefore, the images are aligned considering only the background movement of the living tissue 2.

[0112] In a step 205, the process and display unit 8 is configured to generate a projected image by projection of the pixels of the realigned sharp images, i.e. the image of the output sequence of step 204.

[0113] In a step 205a, the projected values of the pixels of the projected image are minimum intensity values of the pixels of all the sharp images, i.e. of the images of the output sequence of step 204. Preferably, the extremal intensity values of the pixels are minimum intensity values. Therefore, the projection is extracting each point in the sequence where at least one erythrocyte (for example a black dot) has passed once during the time of acquisition by the camera 5. All the positions of all erythrocytes of all the sharp images are rendered in the projected image. The most visible erythrocyte is kept for generating the image. As a result, the “minima image” (i.e. the output image of step 205a) presents the lumens of the capillary blood vessels, as if imaged by a homogeneous light from the inside the blood vessel. The output sequence of step 204 can be projected by using the function “Zproject” from the software ImageJ, and by choosing the option “minimum intensity”. As an alternative, the process can have a step of image value inversion before step 205, and in step 205a, the projected values of the pixels of the projected image are maximum intensity values of the pixels of all the sharp images.

[0114] In a step 205b, the projected values of the pixels of the projected image are the average intensity values of the pixels of all the sharp images, i.e. of the images of the output sequence of step 204. In step 205b, the average red blood cell count passed at each point over time is measured. Therefore, the output image after step 205b corresponds to the flowrate of erythrocytes.

[0115] In a step 205c, the projected values of the pixels of the projected image are the maximal intensity values of the pixels of all the sharp images, i.e. of the images of the output sequence of step 204. In step 205c, the maximal value returns a “white pixel” if the erythrocytes have moved in the region of this pixel since there will be at least one image in which the pixel will be transiently white, this white level representing the light level of the plasma between two flowing erythrocytes. Conversely, if an erythrocyte is always static at the position of this pixel, the maximal value will be equal to the “black” level of this erythrocyte, since, being immobile, it will be present in all images. Therefore, the output image after step 205c corresponds to zones where the flow is stagnant (zero flow).

[0116] Therefore, the imaging of the erythrocyte flow in capillary blood vessels by the device 1 is cheap, does not require fluorescent labelling of cells, does not require injection of fluorescent dies such as FITC dextran and does not require fixation of the living tissue 2. Moreover, the magnitudes of the flow can be measured qualitatively and quantitatively. The image is formed from sources dispersed inside the vessel lumen: therefore, a volume rendering of the vascular surface profile can be imaged by the device 1. Images are obtained by following individual erythrocytes passing over a homogeneous white field. The resolution of the method is for example good enough to provide images of capillaries at magnifications 1× to 9× with a binocular.

[0117] Preferably, imaging can be time lapsed. The device 1 can comprise a control unit 9 coupled to the camera 5. The control unit 9 is configured to control the camera so as to acquire a plurality of sequences along time. An image can be obtained from each sequence, following at least the steps 201, 202, 203, 204 and 205a and/or 201, 202, 203, 204 and 205b or 205c. A video of the obtained images can be made, so as to display the evolution of the capillary blood vessels. Preferably, each sequence is acquired during a given time duration, and the control unit 9 is configured such as a given length of time separates each sequence. Notably, the time lapse recording can last for hours or days, while recording can last preferably less than 1 minute. Each recording of a sequence of images can, for example, be separated by a time length of 5 minutes.

EXAMPLES

[0118] Referring to FIG. 3, FIG. 4 and FIG. 5, images of a CAM vasculature at day 10 of development of the embryo, are obtained by the device 1. The images are taken with a low resolution binocular, allowing imaging in wide field (1× in FIG. 3) but also at higher level of magnification (2× in FIG. 4 and 4× in FIG. 5). FIG. 3, FIG. 4 and FIG. 5 illustrate the interlace structure of the capillaries.

[0119] Referring to FIG. 6, the terminal arterioles can be imaged by the device 1. Vertical chimneys can be observed (illustrated by arrows in FIG. 6), regularly spaced along the arterioles, from which the flow is higher (and not along veins). These chimneys are visible by hotspots corresponding to erythrocytes flowing towards the plane of view of FIG. 6. They appear bright because of the stacking effect of the red cells which increases the contrast. However, as one approaches the distal end of the arteriole, there exists a flat area devoid of vertical chimneys (illustrated by the brackets in FIG. 6).

[0120] Referring to FIG. 7, hotspots point can be visible, for example up to the 14.sup.th day of development of the embryo. The hotspots points are indicated by the white arrows. The hotspot points correspond to capillaries presenting a direction out of the plane of observation of FIG. 7. As the CAM develops, the interlace becomes thinner, with smaller capillary anastomoses.

[0121] Referring to FIG. 21, capillaries of very small size, for example having a diameter smaller than 8 micrometers can be imaged with the device 1.

[0122] Despite ample embryo movements at later stages (14.sup.th day of development of the embryo, FIG. 8), the capillary lattice can be imaged by the device 1, and the vessels at the upper levels of the hierarchy can also be imaged by filming the embryo at the periphery of the plastic cup 10 where movements are smaller, or by carefully selecting a favorable area elsewhere.

[0123] Referring to FIG. 9 and to FIG. 10, arterioles exhibit a terminal segment which is distinctly different from the proximal segment. The arterial pattern shows the presence of vertical “chimneys” or “springs” of flow identified by the presence of hotspots along the arterioles. The hotspots provide flow sources located proximally with respect to the arteriolar hierarchy. The veins do not navigate towards the tip of arteries where the flat distal areas are found (illustrated by arrows in FIG. 10). Referring to FIG. 10, a magnification shows that the flat distal areas are wider than the more proximal part of the arteries. The distal segment often has a “goose leg” widening (illustrated in FIG. 10 by arrowheads).

[0124] Referring to FIG. 11, FIG. 12, FIG. 13 and FIG. 14, the optical profile of vessels can be calculated and plotted from an image of the device 1. The terminal parts of the arterioles appear flat (graphs D in FIG. 11, FIG. 12, FIG. 13 and FIG. 14), while more proximal vessels appear round (graphs D in FIG. 11, FIG. 12, FIG. 13 and FIG. 14). The white lines correspond to the locations where the profiles were extracted, D stands for Distal and P for Proximal. This shows that the terminal part of the arteriole is flattened on the underneath surface of the ectoderm on which it creeps (N=30 arterioles were processed giving similar results).

[0125] Referring to FIG. 15, the device 1 can image 3D collaterals pointing upwards attracting veins. The vertical chimneys along the arterioles serve as flow sources attracting hemodynamically the veins proximally. FIG. 15 corresponds to a day 7 embryo, in which the vessels are quite coarse.

[0126] FIG. 16 corresponds to the vascular interlace at day 10. The capillary plexus is oriented at day 10 along the path starting at the vertical chimneys and flowing towards the presumptive vein. Careful inspection allows one to distinguish the swollen area under the distal flattened segment of the arteriole. The brackets show the flat distal segments. A dimmer area corresponding to the flattened plexus is imaged around the two segments close to the brackets.

[0127] FIG. 17 illustrates an image acquired by the device 1, corresponding to the output of step 205a, wherein the generation of projected image is processed with minimal intensity values of the pixels of the realigned image.

[0128] FIG. 18 illustrates an image acquired by the device 1, corresponding to the output of step 205b, wherein the generation of projected image is processed with average intensity values of the pixels of the realigned image. An area presenting a lower flowrate is illustrated by a star in FIG. 17, along the “flat” distal part of the vessel. The veins develop towards the vertical chimneys located more proximally along the arteries, illustrated by white arrows. The device 1 images typical interdigitating artery and veins, in both average mode and minimal intensity mode.

[0129] FIG. 19 illustrates an image of the input sequence acquired by the camera 5 before step 201. In comparison, FIG. 20 illustrates an image obtained by the device 1 after step 205a, corresponding to the same field of view than in FIG. 19.

[0130] FIG. 20 illustrates an image of the input sequence at the magnification ×8, showing a scale bar of 40 micrometers, and the typical dimension of capillaries at this stage, around and below 8 micrometers.

[0131] In reference to FIG. 21, the depth of at least a capillary blood vessel can be calculated from the intensity of the projected image generated by projection of the pixels having an extremal value in step 205a, so as to deduce the cross section of the capillary blood vessels. In a step 151, the projected image of step 205b is calculated, for example from the acquisition of 300 images at 15 Hz. The projected image is inverted to get a gray level view of the absorption. In step 152, a section is chosen, illustrated by a white line in FIG. 14. The absorption across the vessel in the vertical direction is proportional to local vessel depth. In step 153, a profile of absorption perpendicularly to the vessel is extracted from an intensity measurement of the projected image, at some straight location in between collaterals where a tubular segment is found. The profile of absorption has a height proportional to vessel depth. In a step 154, the measured halve is symmetrized. In a step 155, the vessel cross section is reconstructed by assembling both halves, by assuming a bilateral symmetry. Larger vessels which have higher pressure and are obviously almost cylindrical can be used for calibration. We assume in this set up that the optical refraction index of the plasma is identical to the optical refraction index of the tissue. This amounts to neglecting the cylindrical aberration due to the cylinder/plan diopter. For adjusted indices, the optical path across a cylinder embedded in an image behaves as for a flat diopter, and there is no refractive distortion.

[0132] FIG. 22 is an image of the CAM vasculature acquired and processed by the device. The scale bar corresponds to 40 μm. The image is acquired at a magnification ×8.