Abstract: An apparatus includes a memory and a circuit. The memory may be configured to store data. The circuit may be configured to (i) analyze a warp field to determine a current mode of a plurality of read modes that reduces a memory bandwidth, (ii) read one or more rectangular regions of an input image containing data from the memory based on the current mode, and (iii) generate an output image by filtering the data received from the memory with a warp field.

Abstract: An electronic information board apparatus includes: a guide generating unit configured to display a handwriting region on a screen; a coordinate detecting unit configured to detect coordinates of an indication body moving in the handwriting region on the screen; an image drawing unit configured to generate a stroke image based on the coordinates and display the generated stroke image in the handwriting region on a first layer of the screen; a character recognizing unit configured to execute character recognition based on a hand-written image that is hand-written inside the handwriting region and outputs text data; and a display superimposing unit configured to display the text data acquired from the character recognizing unit at a position that is approximately the same as that of the hand-written image that is hand-written inside the handwriting region on the screen, and on a second layer of the screen different from the first layer.

Abstract: A facial image that a user desires to look for from among a plurality of detected facial image displayed in a first facial image display area is clicked, and the clicked facial image is displayed in a third facial image display area. A facial image that is similar to the facial image displayed in the third facial image display area and is suitable for printing is looked for among from the plurality of detected facial image displayed in the first facial image display area, and displayed in a second facial image display area. The user clicks the facial image of the person which the user desires to look for from among the facial images displayed in the second facial image display area. The clicked facial image is displayed in the third facial image display area.

Abstract: The technology disclosed relates to constructing a convolutional neural network-based classifier for variant classification. In particular, it relates to training a convolutional neural network-based classifier on training data using a backpropagation-based gradient update technique that progressively match outputs of the convolutional neutral network-based classifier with corresponding ground truth labels. The convolutional neural network-based classifier comprises groups of residual blocks, each group of residual blocks is parameterized by a number of convolution filters in the residual blocks, a convolution window size of the residual blocks, and an atrous convolution rate of the residual blocks, the size of convolution window varies between groups of residual blocks, the atrous convolution rate varies between groups of residual blocks. The training data includes benign training examples and pathogenic training examples of translated sequence pairs generated from benign variants and pathogenic variants.

Abstract: An example apparatus for calibrating texture cameras includes an image receiver to receive a depth image from a depth camera and a color image from a texture camera. The apparatus also includes a feature extractor to extract features from the depth image and the color image. The apparatus further includes a feature tester to detect that the extracted features from the depth image and the color image exceed a quality threshold. The apparatus includes a misalignment detector to detect a misalignment between the extracted features from depth image and the extracted features from color image exceeds a misalignment threshold. The apparatus also further includes a calibrator to modify calibration parameters for the texture camera to reduce the detected misalignment between the extracted features from the depth image and the extracted features from the color image below a misalignment threshold.

Abstract: The technology disclosed includes systems and methods to reduce overfitting of neural network-implemented models that process sequences of amino acids and accompanying position frequency matrices. The system generates supplemental training example sequence pairs, labelled benign, that include a start location, through a target amino acid location, to an end location. A supplemental sequence pair supplements a pathogenic or benign missense training example sequence pair. It has identical amino acids in a reference and an alternate sequence of amino acids. The system includes logic to input with each supplemental sequence pair a supplemental training position frequency matrix (PFM) that is identical to the PFM of the benign or pathogenic missense at the matching start and end location. The system includes logic to attenuate the training influence of the training PFMs during training the neural network-implemented model by including supplemental training example PFMs in the training data.

Abstract: An image-processing device is provided with: a pattern-matching portion that performs pattern matching between a first image and a second image acquired before the first image and that outputs a motion vector between the images and an evaluation value that indicates the accuracy of the motion vector; a recursive noise-reduction portion that outputs a noise reduction image, which is the first image to which noise reduction has been applied, and that performs noise reduction in which an image based on the first image and an image based on the noise reduction image formed in the past are mixed in accordance with a predetermined recursion amount; and a controlling portion that controls a recursion amount in the recursive noise-reduction portion on the basis of the motion vector and the evaluation value output by the pattern-matching portion and a gain that is applied to the first image to adjust a brightness thereof.

Abstract: Described is an assistance device for providing imaging support to an operating surgeon during a surgical procedure involving at least one medical instrument. The assistance device comprises a camera, a display unit, a manipulator coupled to the camera, a manipulator controller and an image processing unit. The image processing unit includes an instrument detection module for detecting at least one target structure that represents the instrument being used in the frame in question by identifying a predetermined distinguishing feature, and for extracting position information that indicates the position of the target structure in the frame. The instrument detection module identifies an image segment as the predetermined distinguishing feature, said image segment being characterized by a color saturation that is equal to or less than a predefined color saturation and by a contour line that delimits the image segment and has at least one rectilinear section.

Abstract: Methods, devices, and apparatus, including computer programs encoded on a computer storage medium for reconstructing image are provided. In one aspect, a method of reconstructing image includes obtaining scanning data for a subject in a continuous incremental scanning of medical equipment including real crystals for detection, associating each of the real crystals with one or more virtual crystals in a virtual scanning system, determining delay random coincidence data of two virtual crystals connected by a response line in the virtual scanning system, obtaining random coincidence data by denoising the delay random coincidence data based on crystal receiving efficiency for each of the real crystals, and reconstructing an image with the scanning data by taking the random coincidence data into account.

Abstract: One example method to reduce image artifacts, which may include obtaining measured projection data acquired using an imaging system. The measured projection data is associated with a target object and an artifact source within a radiation field of the imaging system. The method may also include generating virtual projection data associated with the artifact source by forward projecting a model representing one or more physical properties of the artifact source. The method may further include generating corrected projection data based on the measured projection data and the virtual projection data; and reconstructing the corrected projection data into reconstructed volume image data to reduce image artifacts caused by the artifact source.

Abstract: The present invention relates to a method and system for providing ‘diagnosis-aiding information’, the method and system being capable of diagnosing cancer and the like of a patient by using medical imaging.

Abstract: The present invention provides a method for extracting and classifying features of hyperspectral remote sensing image, including: an sampling step, a binarizing step, a coding step, a statistical calculating step, a concatenating step, and a classifying step. The present invention further provides a system for extracting and classifying features of hyperspectral remote sensing image. The technical solution provided by the present invention can make full use of the contextual relationship between the spectral domain and the spatial domain in a hyperspectral remote sensing image by extending two-dimensional LBPs into three-dimensional LBPs, and has good robustness to noise by introducing a relaxation threshold discrimination operation.

Abstract: A Dynamic Vision Sensor (DVS) pose-estimation system includes a DVS, a transformation estimator, an inertial measurement unit (IMU) and a camera-pose estimator based on sensor fusion. The DVS detects DVS events and shapes frames based on a number of accumulated DVS events. The transformation estimator estimates a 3D transformation of the DVS camera based on an estimated depth and matches confidence-level values within a camera-projection model such that at least one of a plurality of DVS events detected during a first frame corresponds to a DVS event detected during a second subsequent frame. The IMU detects inertial movements of the DVS with respect to world coordinates between the first and second frames. The camera-pose estimator combines information from a change in a pose of the camera-projection model between the first frame and the second frame based on the estimated transformation and the detected inertial movements of the DVS.

Abstract: A method for identifying skin region in an image includes: obtaining a target image; and for each of a plurality of pixels in the target image, calculating a probability that the pixel corresponds to skin captured in the image, wherein calculating the probability includes: calculating a first probability of said each pixel in the target image in a first color space the first probability being a probability that the pixel is skin in the first color space; calculating a second probability of said each pixel in the target image in a second color space, the second probability being a probability that the pixel is skin in the second color space; and determining a combined probability that said each pixel in the target image is skin, the combined probability that the pixel is skin being an arithmetic mean value of the first probability and the second probability of the pixel.

Abstract: An image processing pipeline may perform temporal filtering on independent color channels in image data. A filter weight may be determined for a given pixel received at a temporal filter. The filter weight may be determined for blending a value of a channel in a full color encoding of the given pixel with a value of the same channel for a corresponding pixel in a previously filtered reference image frame. In some embodiments, the filtering strength for the channel may be determined independent from the filtering strength of another channel in the full color encoding of the given pixel. Spatial filtering may be applied to a filtered version of the given pixel prior to storing the given pixel as part of a new reference image frame.

Abstract: A system for detecting and responding to an intruding camera. The system includes an electronic media display device having a screen configured to display content, a sensor, and a processing circuit. The processing circuit is configured to obtain information from the sensor, analyze the information to determine a presence of a camera, and edit any displayed content in response to the presence of the camera.

Abstract: Methods and apparatus automatically classify intravascular plaque using features extracted from intravascular optical coherence tomography (IVOCT) imagery. One example apparatus includes an image acquisition circuit that accesses a set of IVOCT images, a pre-processing circuit that generates a blood vessel mask based on the IVOCT images, a feature extraction circuit that defines a three dimensional (3D) volume of interest centered on a location in a member of the set of IVOCT images, a classification circuit that generates a classification based on a probability that a voxel represents a type of plaque, and a visualization circuit that provides a visualization, substantially in real time, of a member of the set of IVOCT images and the classification, where the visualization includes a sector classification image, a labeled image, or a 3D visualization. A prognosis or treatment plan may be provided based on the visualization or the classification.

Abstract: Techniques for high dynamic range image processing are presented. Base layer data, first checksum parameter, and residual ratio data for a high dynamic range (HDR) image are each received. A second checksum parameter is computed for the base layer data based upon the first SOF after the last APP11 marker segment and includes all following bytes up to and including the EOI marker. The first and second checksum parameters are compared to determine if base layer has been altered.

Abstract: The technology disclosed relates to constructing a convolutional neural network-based classifier for variant classification. In particular, it relates to training a convolutional neural network-based classifier on training data using a backpropagation-based gradient update technique that progressively match outputs of the convolutional network network-based classifier with corresponding ground truth labels. The convolutional neural network-based classifier comprises groups of residual blocks, each group of residual blocks is parameterized by a number of convolution filters in the residual blocks, a convolution window size of the residual blocks, and an atrous convolution rate of the residual blocks, the size of convolution window varies between groups of residual blocks, the atrous convolution rate varies between groups of residual blocks. The training data includes benign training examples and pathogenic training examples of translated sequence pairs generated from benign variants and pathogenic variants.

Abstract: Provided is a medical image processing apparatus including: a data acquisition unit configured to respectively acquire a plurality of medical images representing an object including at least one target at a plurality of different time points; and an image processor configured to generate, based on the acquired plurality of medical images, a diagnostic image showing a degree of change that has occurred in the at least one target over the plurality of different time points.