Disclosed are various embodiments for automated pill identification using lighting devices and machine learning routines. A computing device may selectively control illumination of a pill provided at an imaging position by a pill dispensing system. The computing device may direct an imaging device to capture image data of the pill during illumination of the pill. Also, the computing device may generate a digital fingerprint of the pill and determine an identity of the pill based at least in part on a comparison of the digital fingerprint to a digital fingerprint library. A machine learning routine may be applied to improve future detection of the identity of the pill.

A system and method for scoring trained probes for use in analyzing one or more candidate poses of a runtime image is provided. A set of probes with location and gradient direction based on a trained model are applied to one or more candidate poses based upon a runtime image. The applied probes each respectively include a discrete set of position offsets with respect to the gradient direction thereof. A match score is computed for each of the probes, which includes estimating a best match position for each of the probes respectively relative to one of the offsets thereof, and generating a set of individual probe scores for each of the probes, respectively at the estimated best match position.

The present disclosure discloses an image processing method and device. The image processing method includes: dividing a detection image into a plurality of first subregions, dividing a template image into a plurality of second subregions, calculating a principal rotation direction of each first subregion with respect to the corresponding second subregion; and calculating a principal rotation direction of the detection image according to the principal rotation directions of the plurality of first subregions.

An approach is provided for a vertex-based evaluation of polygon similarity. The approach, for instance, involves processing, by a computer vision system, an image to generate a first set of vertices of a first polygon representing an object depicted in the image. The approach also involves for each vertex in the first set of vertices, determining a closest vertex in a second set of vertices of a second polygon, and determining a distance between said each vertex in the first set of vertices and the closest vertex in the second set of vertices. The approach further involves calculating a polygon similarity of the first polygon with respect to the second polygon based on a total of the distance determined for said each vertex in the first set of vertices normalized to a number of vertices in the first set of vertices.

A controller/application analyzes image data from a camera to determine the distance to an object in an image based on the size of the object in the image and based on a known focal length of a camera that captured the image and based on a known dimension of the actual object. The known dimension of the object may be retrieved from a database that is indexed according to outline shape, color, markings, contour, or other similar features or characteristics. The distance determined from analysis of the image and objects therein may be used to calibrate, or to verify the calibration of, complex distance determining systems that may include LIDAR. Object distance determinations in different image frames, whether to the same or different object, taken a known time apart may be used to determine speed of the camera that took the images, or speed of a vehicle associated with camera.

First, the data of a moving image that is captured is read for each frame, and whether to start tracking is determined based on the presence or absence of a target object (S20, S22). An edge image of the image frame is created after it is determined to start tracking (S24). Meanwhile, a particle is distributed in a space of a coefficient set for each control point sequence when the control point sequence of a B-spline curve representing the shape of the target object is represented in a linear combination of control point sequence of a B-spline curve representing a plurality of reference shapes that are made available in advance (S26). A particle is also distributed in the space of a shape-space vector (S28), the likelihood of each particle is observed, and the probability density distribution is acquired (S30). A curve obtained by weighting parameters by the probability density distribution and then averaging the weighted parameters is created as a tracking result (S32).

An image processing method can suppress detection accuracy of a detection target object from being lowered even if the detection target object has a different surface condition because of the influence of various kinds of noise. The image processing method includes the following operations of generating a captured model edge image by executing edge extraction processing on a captured model image acquired by capturing a detection target object, executing pattern matching of the captured model edge image and a model edge image, calculating similarity at respective edge points in the model edge image in the pattern matching of the captured model edge image and the model edge image, selecting an edge point to be eliminated based on the similarity from among the respective edge points in the model edge image, and generating an edge image acquired by eliminating the selected edge point as a final model edge image.

A shape similarity metric can be provided that indicates how similar two or more shapes are. A difference between a union of the shapes and an intersection of the shapes can be used to determine the similarity metric. The shape similarity metric can provide an average distance between the shapes. Different processes for determining shapes can be evaluated for accuracy based on the shape similarity metric. New or alternative shape-determining processes can be compared for accuracy against other shape-determining processes including reference shape-determining processes. Shape similarity metrics can be determined for two-dimensional shapes and three-dimensional shapes.

Registration of a surgical image acquisition device (e.g. an endoscope) using preoperative and live contour signatures of an anatomical object is described. A control unit includes a processor configured to compare the real-time contour signature to the database of preoperative contour signatures of the anatomical object to generate a group of potential contour signature matches for selection of a final contour match. Registration of an image acquisition device to the surgical site is realized based upon an orientation corresponding to the selected final contour signature match.

There is provided a method for generating a personalized Head Related Transfer Function (HRTF). The method can include capturing an image of an ear using a portable device, auto-scaling the captured image to determine physical geometries of the ear and obtaining a personalized HRTF based on the determined physical geometries of the ear.