Abstract: Combined color and depth data for a field of view is received. Thereafter, using at least one bounding polygon algorithm, at least one proposed bounding polygon is defined for the field of view. It can then be determined, using a binary classifier having at least one machine learning model trained using a plurality of images of known objects, whether each proposed bounding polygon encapsulates an object. The image data within each bounding polygon that is determined to encapsulate an object can then be provided to a first object classifier having at least one machine learning model trained using a plurality of images of known objects, to classify the object encapsulated within the respective bounding polygon.

Abstract: A diagnosis support apparatus for diagnosing a lesion based on a captured image composed of a multi-valued image, the apparatus includes a dermoscope-equipped image capturing device which captures the captured image; a storage device which stores the captured image; a display which displays the captured image stored in the storage device; and a processor which, under control of a stored program, processes the captured image stored in the storage device.

Abstract: According to one embodiment, a method of image analysis is provided. The method includes binning image data into a plurality of sinogram frames, identifying a plurality of initial stationary frames by applying a first analysis technique on the plurality of binned sinogram frames, extracting a plurality of first statistical parameters applying a second analysis technique on the plurality of binned sinogram frames, combining the plurality of first statistical parameters with boundaries of plurality of initial stationary frames to generate a presentation of a joint analysis combining at least some of the plurality of the first statistical parameters and at least some of the plurality of the second statistical parameter, identifying a plurality of final stationary frames from the presentation of the joint analysis, independently reconstructing each of the plurality of final stationary frames, and registering each of the plurality of final stationary frames to a first state.

Abstract: The provided is an image processing device, an image processing method, and a program that can solve the above described problems and can normalize the distortion in the a fisheye image without being influenced by receiving the influence of the a fisheye optical system even when the positional relationship between the an optical axis of the a fisheye camera and the a plane to which the an object is grounded does not satisfy a certain conditions.

Abstract: A method of improved quality inspection includes (i) receiving a first command to capture a first image of a surface of a product, (ii) positioning, by actuating a plurality of rotatable joints, a camera at a first position that is substantially adjacent to the surface of the product, and capturing the first image of the surface of the product. The method further includes, after capturing the first image: (i) processing the first image to identify a defect in the first image and a relative location of the defect in the first image, and (ii) determining a second position of the camera in accordance with the first position of the camera and the relative location of the defect in the first image, and (iii) repositioning, by actuating the plurality of rotatable joints, the camera from the first position to the second position. The method further includes capturing the second image.

Abstract: An image processing system includes a processor and optical flow (OF) determination logic for quantifying relative motion of a feature present in a first frame of video and a second frame of video that provide at least one of temporally and spatially ordered images with respect to the two frames of video. The OF determination logic configures the processor to implement performing OF estimation between the first frame and second frame using a pyramidal block matching (PBM) method to generate an initial optical flow (OF) estimate at a base pyramid level having integer pixel resolution, and refining the initial OF estimate using at least one pass of a modified Lucas-Kanade (LK) method to provide a revised OF estimate having fractional pixel resolution.

Abstract: Embodiments described herein provide various examples of a face detection system, based on using a small-scale hardware convolutional neural network (CNN) module configured into a multi-task cascaded CNN. In some embodiments, a subimage-based CNN system can be configured to be equivalent to a large-scale CNN that processes the entire input image without partitioning such that the output of the subimage-based CNN system can be exactly identical to the output of the large-scale CNN. Based on this observation, some embodiments of this patent disclosure make use of the subimage-based CNN system and technique on one or more stages of a cascaded CNN or a multitask cascaded CNN (MTCNN) so that a larger input image to a given stage of the cascaded CNN or the MTCNN can be partitioned into a set of subimages of a smaller size. As a result, each stage of the cascaded CNN or the MTCNN can use the same small-scale hardware CNN module that is associated with a maximum input image size constraint.

Abstract: A system includes a microscope configured to magnify a pathology sample, a camera positioned to record magnified pathology images from the microscope, a pathology database including reference pathology images, and a display configured to show the magnified pathology images. A processing apparatus is coupled to the camera, the pathology database, and the display. The processing apparatus includes instructions that when executed by the processing apparatus cause the system to perform operations. The operations include comparing the magnified pathology images to the reference pathology images included in the pathology database; identifying one or more regions of interest in the magnified pathology images; and alerting, using the display, a user of the microscope to the one or more regions of interest in the magnified pathology images while the pathology sample is being magnified with the microscope.

Abstract: A feature amount, and context information representing a context regarding an image are obtained from the image. The first likelihood indicating the likelihood at which the image is an image of a specific object is obtained from the feature amount. The second likelihood indicating the likelihood at which the image is an image of the specific object is obtained from the context information. Whether the image is an image of the specific object is identified using the first likelihood and the second likelihood.

Abstract: The present disclosure relates to combination of images. A method according to an embodiment comprises: receiving a visual image and an infrared (IR) image of a scene and for a portion of said IR image extracting high spatial frequency content from a corresponding portion of said visual image. The method according to the embodiment further comprises combining said extracted high spatial frequency content from said portion of the visual image with said portion of the IR image, to generate a combined image, wherein the contrast and/or resolution in the portion of the IR image is increased compared to the contrast and/or resolution of said received IR image.

Abstract: Obtaining a good image is important for reading indicia. Often, however, objects are not well aligned with an indicia reader's digital imager. This misalignment may lead to images in which an indicium may appear distorted due to perspective. This perspective distortion makes indicia reading difficult. In general, perspective distortion of an imaged object may be corrected using knowledge of the true size, shape, and/or orientation of the physical object. The present invention embraces the use of a dimensioning system to obtain the 3D information necessary to provide this knowledge. Then, using this knowledge, to transform a distorted image of an indicium into a corrected image of the indicium without distortion. This corrected image may then be processed to read the undistorted indicium.

Abstract: The present invention relates to a method of diagnosis which identifies one or more individual cells and comprises determining at least one of a cell dimension and a cell area, and determining at least one of dark/light cell contrast characteristics, cell area characteristics, cell color characteristics, cell roughness characteristics, distances between cell nuclei and cell convexity. A computer readable medium, a computer apparatus and a diagnostic system is also provided.

Abstract: In generation of a wedgelet pattern, all the patterns are generated through a line segment drawing process and a region filling process. Therefore, there is a problem that a processing amount is large. A moving image decoding device (1) includes: a wedgelet pattern generation section (145T4) that generates a wedge pattern; a rotation pattern generation section (1465) that derives a rotation pattern by rotating the wedge pattern; a pattern storage section (1464) that adds the rotation pattern to a wedge pattern table; and a DC prediction value derivation section (145T3) that derives a prediction value of each region partitioned in the wedge pattern included in the wedge pattern table (145T5).

Abstract: The present invention discloses a method for finding a cell nucleus of a target cell from a cell image, wherein the cell image includes the target cell and at least one variation cell, and the target cell includes cytoplasm and the cell nucleus. The method includes steps of: (a) processing the cell image via an image processor such that the cytoplasm, the cell nucleus and the variation cell have different shades of color; (b) demarcating the outlines of the cytoplasm, the cell nucleus and the variation cell; (c) calculating geometrical reference points of the outlines; (d) calculating the distances from the geometrical reference point of the cytoplasm outline to the geometrical reference point of the cell nucleus outline and to the geometrical reference points of the variation cell outlines; and (e) finding a specific geometrical reference point having a shortest distance to locate a specific outline corresponding to the specific geometrical reference point as the cell nucleus.

Abstract: An optical tracking system comprises a marker part, an image forming part, and a processing part. The marker part includes a pattern having particular information and a first lens which is spaced apart from the pattern and has a first focal length. The image forming part includes a second lens having a second focal length and an image forming unit which is spaced apart from the second lens and forms an image of the pattern by the first lens and the second lens. The processing part determines the posture of the marker part from a coordinate conversion formula between a coordinate on the pattern surface of the pattern and a pixel coordinate on the image of the pattern, and tracks the marker part by using the determined posture of the marker part. Therefore, the present invention can accurately track a marker part by a simpler and easier method.

Abstract: A method of augmenting sight in an individual. The method comprises obtaining an image of a scene using a camera carried by the individual; transmitting the obtained image to a processor carried by the individual; selecting an image modification to be applied to the image by the processor; operating upon the image to create a modified image using either analog or digital imaging techniques, and displaying the modified image on a display device worn by the individual. The invention also relates to an apparatus augmenting sight in an individual. The apparatus comprises a camera, carried by the individual, for obtaining an image of a scene viewed by the individual; a display carried by the individual; an image modification input device carried by the individual; and a processor, carried by the individual. The processor modifies the image and displays the modified image on the display carried by the individual.

Abstract: Back projection voxels smaller in size than forward projection voxels are created. Back projection is performed by the use of computed pixel values and measured pixel values at intersection points between straight lines connecting an X-ray source with the centroids of the back projection voxels and an X-ray detection panel.

Abstract: An image processing device includes a detection unit configured to detect an object from an image captured by each of a plurality of image capturing devices, a determination unit configured to determine whether the object is detected by two image capturing devices or more in the image capturing devices, a first estimation unit configured to estimate, when the object have been detected by the two image capturing devices or more, a three-dimensional position and information on a height of the object based on a detected position where the object is detected in an image captured for each of the two image capturing devices and positions and orientations of the two image capturing devices, and a second estimation unit configured to estimate, when the object have been detected by a single image capturing device, the three-dimensional position of the object.

Abstract: The invention relates to a method for determining an unknown position, i.e. height and/or orientation, of a light source within a locality. The determination is based on a first image of a scene within the locality acquired by a camera in such a manner as to contain a light footprint of light emitted by the light source from the unknown position. The method includes steps of processing the first image to determine one or more characteristics of the at least the portion of the light footprint within the first image, comparing the determined characteristics with one or more corresponding known characteristics of a light footprint of light emitted by the light source from a known position to determine a deviation between the determined and the known characteristics, and determining the unknown position of the light source based on the determined deviation.

Abstract: One example method to improve image quality of projection image data may include obtaining projection image data and channel offset data associated with the projection image data. The channel offset data may be acquired using the flat panel detector and include at least one set of channel offset data values associated with respective channels of the flat panel detector. The method may also include generating channel offset drift data representing one or more variations of the channel offset data from a reference channel offset data. The method may further include generating offset-compensated projection image data by modifying the projection image data based on the channel offset drift data to compensate for the one or more variations of the channel offset data.