Abstract: A determination processing unit determines a disease region in a medical image including an axisymmetric structure. A first determination section of the determination processing unit generates a feature amount map of the medical image from the medical image. A second determination section second inverts the feature amount map with reference to a symmetry axis to generate an inverted feature amount map. A third determination section superimposes the feature amount map and the inverted feature amount map on each other and determines the disease region in the medical image using the feature amount map and the inverted feature amount map superimposed on each other.

Abstract: The invention provides a method for generating a compound ultrasound image. The method includes acquiring and beamforming channel data. Using the beamformed channel data a plurality of images, each image comprising a plurality of pixels, of a region of interest are obtained and an image information metric, wherein the image metric is associated with a pixel of the plurality of pixels, is assessed. The acquiring of the plurality of images and the assessment of the image metric are performed in parallel. For each image of the plurality of images: a per-pixel weighting for each pixel of the plurality of pixels based on the assessment of the image information metric is determined and applied to each pixel of the plurality of pixels. Finally a compound ultrasound image is generated based on the plurality of weighted pixels of the plurality of images.

Abstract: Systems and methods for testing visible chemical reactions of a reagent are provided. In one implementation, the method may include receiving from an image sensor associated with a mobile communications device an image of a reagent with a plurality of colored test reagent pads in proximity to a colorized surface having a plurality of colored reference elements of differing shades. The method further includes using the differing shades of the plurality of colored reference elements to determine local illumination conditions. Thereafter, the method includes using the determined local illumination conditions and an analysis of a depiction of the plurality of colored test reagent pads in the image to determine an extent of a chemical reaction on the reagent. Then the method includes causing the mobile communications device to provide to the user an indication that the testing of the reagent is complete.

Abstract: Among other things, one or more systems and/or techniques for classifying an item disposed within an object are provided herein. A three-dimensional image of the object (e.g., a bag) is segmented into a set of item representations (e.g., laptop, thermos, etc.). An item is identified from the set of item representations based upon item features of the item, such as the laptop that could be used to conceal an item of interest such as an explosive. A region comprising a three-dimensional image of the item is divided into a set of sub-regions (e.g., a first sub-region encompassing a screen, a second sub-region encompassing a motherboard, etc.). The item is classified as a potential first type of item (e.g., an explosive laptop) when any sub-region has a number of voxels, with computed tomography (CT) values within a range of known CT values for a first type of item, exceeding a threshold.

Abstract: An approach is provided for generating a polyline from line segments (e.g., line segments representing objects detected by a computer vision system). The approach involves selecting a line segment from a plurality of line segments. The approach also involves determining a neighboring line segment from among the plurality of line segments. The determined neighboring line segment has a closest distance to the line segment from among the plurality of line segments. The approach further involves merging the line segment and the neighboring line segment into a polyline based on determining that the closest distance is a mutual closest distance between line segment and the neighboring line segment.

Abstract: A system includes a K1 preprocessing module designed to generate at least one intermediate image from an input image using a parameterized internal processing chain and an analysis module to detect a feature or object in the intermediate image. A method to train the system includes feeding a plurality of learning input images to the system, comparing a result provided by the analysis module for each of the learning input images to a learning value, and feeding back a deviation obtained by the comparison to an input preprocessing module and/or adapting parameters of the internal processing chain to reduce the deviation.

Abstract: Techniques are provided to dynamically generate and render an object bounding fence in a mixed-reality scene. Initially, a sparse spatial mapping is accessed. The sparse spatial mapping beneficially includes perimeter edge data describing an object's edge perimeters. A gravity vector is also generated. Based on the perimeter edge data and the gravity vector, two-dimensional (2D) boundaries of the object are determined and a bounding fence mesh of the environment is generated. A virtual object is then rendered, where the virtual object is representative of at least a portion of the bounding fence mesh and visually illustrates a bounding fence around the object.

Abstract: An image processing device includes a processor to detect motion of objects in an observation area on the basis of a captured image. The processor also calculates amounts of each detected motion, and calculates a propagation speed, a propagation direction, or both, of the motion for pulsations of the objects.

Abstract: A method for detecting motion information includes the following steps. First, a pixel array is provided for detecting an image of a measured object located in a first distance range or in a second distance range, and the pixel array includes a plurality of invisible image sensing pixels and a plurality of visible image sensing pixels. Then, image detection is conducted within the first distance range by using the invisible image sensing pixels to output a plurality of invisible images. Next, the image detection is conducted within the second distance range by using the visible image sensing pixels to output a plurality of visible images. Then, the plurality of invisible images and the plurality of visible images are analyzed by using a processing unit, so as to obtain motion information of the measured object. A pixel array for detecting motion information and an image sensor are also provided.

Abstract: A method, an apparatus, a device, and a medium for calibrating a posture of a moving obstacle are provided. The method includes: obtaining a 3D map, the 3D map including first static obstacles; selecting a target frame of data, the target frame of data including second static obstacles and one or more moving obstacles; determining posture information of each of the one or more moving obstacles in a coordinate system of the 3D map; registering the target frame of data with the 3D map; determining posture offset information of the target frame of data in the coordinate system according to a registration result; calibrating the posture information of each of the one or more moving obstacles according to the posture offset information; and adding each of the one or more moving obstacles after the calibrating into the 3D map.

Abstract: A dynamic anatomic atlas is disclosed, comprising static atlas data describing atlas segments and dynamic atlas data comprising information on a dynamic property which information is respectively linked to the atlas segments.

Abstract: A trajectory tracking method is provided for a computer device. The method includes performing motion tracking on head images in a plurality of video frames, to obtain motion trajectories corresponding to the head images; acquiring face images corresponding to the head images in the video frames, to obtain face image sets corresponding to the head images; determining from the face image sets corresponding to the head images, at least two face image sets having same face images; and combining motion trajectories corresponding to the at least two face image sets having same face images, to obtain a final motion trajectory of trajectory tracking.

Abstract: Trajectory prediction may receiving a LiDAR image sequence including a set of LiDAR images and generating a LiDAR map, generating an interaction encoder result by feeding the LiDAR image sequence through an interaction encoder, generating a feature extractor result by feeding the LiDAR map through a feature extractor, generating a relation encoder result by feeding a past trajectory of a detected obstacle from the LiDAR image sequence, the interaction encoder result, and the feature extractor result through a relation encoder, generating an intention estimation result by feeding the relation encoder result through an intention estimator, generating a conditional generative model result by feeding the past trajectory of the detected obstacle, the intention estimation result, and a probability map through a conditional generative model encoder, and generating a trajectory prediction by feeding the relation encoder result, the past trajectory of the detected obstacle, and the conditional generative model result

Abstract: An approach is provided that captures a set of sequential images of an area where there is a selected moving object. Both a keypoint-based (KP-based) matching model and a neural network based (NN-based) matching model are used with the KP-based matching model analyzing most or all of the captured images and the NN-based model being more computational intensive and analyzing a subset of the images. When the KP-based matching model fails to identify the selected object in an image, the NN-based model is used to find the object so that the KP-based matching model can re-establish tracking of the object.

Abstract: An artificial reality system is configured to more accurately and efficiently construct a 3D virtual representation of a real-world environment from a set of 2D images. The system identifies points and/or lines within the images that define a plane along an orientation and then performs a planar sweep along a perpendicular path to identify surfaces in which the plane intersects with multiple points. Points that appear to be in the same plane are “collapsed” into a cohesive plane to conserve processing power by estimating and/or storing parameters for the cohesive plane, rather than all of the individual 3D points. In this way, the system also “averages out” random variation in the planar surface that would otherwise result from random noise in the estimation of the individual 3D points. The system may then generate a 3D map from a constrained alignment of all of the identified planes.

Abstract: Provided is a drawing section that generates a display image from a predetermined viewpoint by using a texture image and a depth image of a first layer and using a texture image and a depth image of a second layer in which a range of information indicated by a pixel value is different from at least one of the texture image or the depth image of the first layer.

Abstract: A method for detecting dimension of box based on depth map includes: receiving a depth map generated by a camera, the depth map corresponds to pixels of an image including a box; performing a coordinate transformation to transform the depth map into camera coordinates of each of the pixels; dividing some of the pixels into plural blocks, each of blocks includes a number of the pixels adjacent to each other; statistically analyzing an average normal vector of each of the blocks according to the camera coordinates of the pixels of each of the blocks; classifying the blocks into plural clusters according to the average normal vector of each of the blocks; performing a plane extraction to obtain edge vectors according to plane formulas of the clusters; obtaining vertexes of the box according to the edge vectors; and obtaining a dimension of the box according the vertexes.

Abstract: An object of the present invention is to estimate the shape of an object easily and with a high accuracy. The present invention is a generation apparatus including: an acquisition unit configured to acquire a plurality of pieces of image data obtained by capturing an object from different directions by a plurality of image capturing apparatuses arranged at different positions, respectively; a first derivation unit configured to derive reliability for each of the image capturing apparatuses based on spatial resolution in the image data; and a generation unit configured to generate three-dimensional shape data representing the shape of the object based on the image data and the reliability.

Abstract: An image processing apparatus includes a representative-distance calculator and a joining processor. The representative-distance calculator generates representative distance values of the basis of a distance image generated from a stereo image including an image of at least one object including a vehicle. The distance image includes distance values of pixels. The joining processor performs a joining process of joining a first image region and a second image region that are defined based on the image of the at least one object and disposed apart from each other in the stereo image. The joining processor performs a determining process of determining whether the first image region and the second image region each include an image of a side face of the vehicle. The joining processor performs the joining process when the first image region and the second image region each include the image of the side face of the vehicle.

Abstract: A camera apparatus is provided for detecting a stream of objects moving relative to the camera apparatus having a plurality of individual cameras that each have an image sensor for recording frames, wherein the frames overlap one another in part, having an evaluation unit for compiling frames, and having a geometry detection sensor for detecting geometrical data of the objects. The evaluation unit is here configured to generate an object image assembled from frames of an individual object of the stream of objects, with the selection of the participating frames and/or the assembly taking place on the basis of the geometrical data.

Abstract: A cargo-sensing unit including: an image sensor; at least one processor; and a memory having stored thereon computer program code that, when executed by the processor, controls the at least one processor to: instruct the image sensor to capture an image of a cargo space within a cargo container; compare the captured image to a baseline image of an empty cargo container; and determine, based on the comparison between the captured image and the baseline image, a cargo space utilization estimate.

Abstract: To precisely specify lane regions that are present in an image. In a situation in which a lane specification device an imaging device, and a swimming pool are present, a processor in the lane specification device according to the present disclosure specifies lane regions that are present in an image on the basis of a pattern in which information obtained on the basis of outlines of objects in the image captured by the imaging device appears in the image. In this manner, it is possible to precisely specify the lane regions that are present in the image.

Abstract: A camera is arranged on a transmitter or receiver mount configured to provide a transmitter or receiver with a field of regard. Image data of the field of regard is captured by the camera. A location of an obscuration within the field of regard from the image data is determined from the image data. A map of obscurations within the field of regard is generated based upon the image data and the location of the obscuration within the field of regard.

Abstract: Disclosed are methods, apparatus and systems for gesture recognition based on neural network processing. One exemplary method for identifying a gesture communicated by a subject includes receiving a plurality of images associated with the gesture, providing the plurality of images to a first 3-dimensional convolutional neural network (3D CNN) and a second 3D CNN, where the first 3D CNN is operable to produce motion information, where the second 3D CNN is operable to produce pose and color information, and where the first 3D CNN is operable to implement an optical flow algorithm to detect the gesture, fusing the motion information and the pose and color information to produce an identification of the gesture, and determining whether the identification corresponds to a singular gesture across the plurality of images using a recurrent neural network that comprises one or more long short-term memory units.

Abstract: This disclosure presents systems and methods to facilitate interaction by one or more participants with content presented across multiple distinct physical locations. A current distinct physical location of a participant may be determined. In response to determining the current distinct physical location in which the participant is located, operation of one or more content devices physically present in the current distinct physical location may be effectuated.

Abstract: This application provides a method and a device for determining an area of interest based on geolocation data. The method includes: dividing an area on a map into a plurality of blocks; for each block in the area, determining the number of terminals within the block in a duration as a value of the block in the duration; determining a plurality of anchor points based on the blocks having values greater than a threshold; and obtaining a curve based on some or all of the anchor points, wherein an area bounded by the curve is an area of interest.

Abstract: A method for determining a coordinate of a feature point of an object in a 3D space comprises: arranging at least one dedicated pointer in the 3D space in a pre-determined relation to the feature point of the object in the 3D space, wherein each dedicated pointer has at least one visible feature in line of sight of the camera; capturing at least one image by using the camera; performing image feature detection on the at least one captured image to determining a coordinate of the respective at least one visible feature of each dedicated pointer; and determining the coordinate of the feature point of the object in the 3D space based on the determined coordinate of the respective at least one visible feature of each dedicated pointer.

Abstract: A method for accurately determining a positioning of an automated guided vehicle relative to a substrate includes scanning a substrate surface of the substrate to produce two or more two-dimensional images, with each respective one of the produced two-dimensional images is associated with a known portion of the substrate surface. The method further includes storing each one of the produced images. The method further includes positioning the automated guided vehicle relative to the substrate surface, and scanning the substrate surface corresponding to the positioned automated guide vehicle to produce a supplemental two-dimensional image. The supplemental image is compared to each of the stored two or more images to accurately determine the position of the vehicle relative to the substrate surface at the position. This information can also be used to drive the automated guided vehicle from the determined position to another known position relative to the substrate surface.

Abstract: A depth-image processing device includes: a memory; and a processor coupled to the memory and configured to: generate, cased on a synthetic model in which a three-dimensional model of a human body and a three-dimensional model of an object are combined, a plurality of learning images in which a depth image that indicates a distance from a reference position to respective positions on the human body or to respective positions on the object, and a part image to identify any one of respective parts of the human body and a cart of the object are associated with each other, and learn an identifier in which a feature of the depth image and any one of a part of the human body and a part of the object are associated with each other, based on the learning images.

Abstract: A vehicle can use an image sensor to both detect objects and determine depth data associated with the environment the vehicle is traversing. The vehicle can capture image data and lidar data using the various sensors. The image data can be provided to a machine-learned model trained to output depth data of an environment. Such models may be trained, for example, by using lidar data and/or three-dimensional map data associated with a region in which training images and/or lidar data were captured as ground truth data. The autonomous vehicle can further process the depth data and generate additional data including localization data, three-dimensional bounding boxes, and relative depth data and use the depth data and/or the additional data to autonomously traverse the environment, provide calibration/validation for vehicle sensors, and the like.

Abstract: According to an exemplary embodiment of the present disclosure, a computer program stored in a computer readable storage medium is disclosed. The computer program causes one or more processors to perform operations below for performing calibration when the one or more processors are executed, and the operations may include: an operation of inputting an image to an object detection model and performing object detection by using the object detection model; an operation of acquiring bounding box information on the detected object; and an operation of performing calibration on an image acquisition device based on at least a part of the bounding box information.

Abstract: The present teaching relates to method, system, medium, and implementations for trailer pose estimation. At least a dimension of a trailer connected to a vehicle is obtained where the vehicle and the trailer have respective opposing surfaces facing one another. A first set of feature points is identified from a first image of a target present on one the opposing surfaces acquired at a first time instant by a sensor deployed on the other the opposing surfaces. A second set of corresponding feature points is identified from a second image of the target acquired at a subsequent time instant by the sensor. Based on the first and second sets of feature points, a first transformation is determined and used to compute a second transformation that the trailer undergoes from the first to the second time instants. A pose of the trailer at the second time instant is estimated based on the second transformation and the dimension.

Abstract: In an approach for chemical detection for a water source, a processor receives, from a network device of a plurality of network devices, data, wherein the data includes at least one of image data, video data, chemical sensor data, and biosensor data. A processor updates a predictive model with the data. A processor receives a table of possible chemical compositions from the predictive model. A processor determines to send an alert based on the table of possible chemical compositions. A processor sends the alert to a user device.

Abstract: An image processing system according to an aspect of the invention includes: an optical sensor including: an optical member configured to refract light from a subject, the optical member including a plurality of transmissive units that have different optical properties, thus forming different refraction patterns of light that passes therethrough; and a plurality of image sensors configured to receive light that has passed through the transmissive units and to form captured images in which the subject appears; and an information processing apparatus configured to calculate attribute information representing an attribute of the subject by inputting the captured images obtained by the image sensors to a learning device that has learned the attribute of the subject.

Abstract: A three-dimensional data encoding method includes: extracting, from first three-dimensional data, second three-dimensional data having an amount of a feature greater than or equal to a threshold; and encoding the second three-dimensional data to generate first encoded three-dimensional data. For example, the three-dimensional data encoding method may further include encoding the first three-dimensional data to generate the second encoded three-dimensional data.

Abstract: The present invention relates to an image encoding/decoding method and apparatus, A method of decoding an image according to an exemplary embodiment of the present invention comprises: decoding transformation information; creating at least one transformed reference picture by applying the transformation information to at least one reference picture organized in a reference picture list; inserting the at least one transformed reference picture into the reference picture list; and performing motion compensation by using the reference picture list.

Abstract: There is provided an information processing apparatus and a method that allow for suppression of a decrease in encoding efficiency. Correlation information is generated that results from encoding of voxel data resulting from quantization of point cloud data with use of correlation of a distribution pattern of values of the voxel data; the generated correlation information is encoded; and a bit stream including the correlation information is generated. The present disclosure is applicable to an information processing apparatus, an image processing apparatus, an electronic device, an information processing method, a program, or the like, for example.

Abstract: A computer-implemented method includes receiving a base visualization having first data in a first set of channels, where each channel in the first set of channels is associated with a respective range in the base visualization. It is detected that the respective ranges of the first set of channels fall outside a perceptual bandwidth of a first user. The base visualization is automatically transformed to a second visualization, based on the perceptual bandwidth of the first user. The second visualization includes second data in a second set of channels, where each channel in the second set of channels is associated with a respective range in the second visualization. The respective ranges of the second set of channels fall within the perceptual bandwidth of the first user.

Abstract: An interactive palette interface includes a color picker for digital paint applications. A user can create, modify and select colors for creating digital artwork using the interactive palette interface. The interactive palette interface includes a mixing dish in which colors can be added, removed and rearranged to blend together to create gradients and gamuts. The mixing dish is a digital simulation of a physical palette on which an artist adds and mixes various colors of paint before applying the paint to the artwork. Color blobs, which are logical groups of pixels in the mixing dish, can be spatially rearranged and scaled by a user to create and explore different combinations of colors. The color, position and size of each blob influences the color of other pixels in the mixing dish. Edits to the mixing dish are non-destructive, and an infinite history of color combinations is preserved.

Abstract: Systems and methods for transforming grayscale images into color images using deep neural networks are described. One of the systems include one or more computers and one or more storage devices storing instructions that, when executed by one or more computers, cause the one or more computers to implement a coloring neural network, a refinement neural network, and a subsystem. The coloring neural network is configured to receive a first grayscale image having a first resolution and to process the first grayscale image to generate a first color image having a second resolution lower than the first resolution. The subsystem processes the first color image to generate a set of intermediate image outputs. The refinement neural network is configured to receive the set intermediate image outputs, and to process the set of intermediate image outputs to generate a second color image having a third resolution higher than the second resolution.

Abstract: In implementations of weighted color palette generation, one or more computing devices implement a generation system which receives input data including an input color palette. A first machine learning model receives the input color palette and generates an unweighted color palette based on the input color palette. A second machine learning model receives the generated unweighted color palette and generates a weighted color palette based on the generated unweighted color palette. The generation system renders the weighted color palette in a user interface.

Abstract: Image processing systems and methods are disclosed, including an image processing system comprising a computer running image processing software causing the computer to: divide an oblique aerial image into a plurality of sections, choose reference aerial image(s), having a consistent color distribution, for a first section and a second section; create a color-balancing transformation for the first and second sections of the oblique aerial image such that the first color distribution of the first section matches the consistent color distribution of the chosen reference aerial image and the second color distribution of the second section matches the consistent color distribution of the chosen reference aerial image; color-balance pixel(s) in the first and section sections of the oblique aerial image, such that at least one color-balancing transformation of the first and second sections matches the consistent color distribution of the reference aerial image(s).

Abstract: The present disclosure provides a method for processing projection data. The method may include obtaining a first image generated by performing a first scan to a subject by a first imaging device; determining first projection data based on the first image, the first projection data corresponding to a first area of the subject; obtaining second projection data by performing a second scan of the subject using a second imaging device, the second projection data corresponding to a second area of the subject, the first area at least partially overlapping with the second area in an overlapping area; determining registered first projection data by registering the first projection data to the second projection data with respect to the overlapping area; determining scatter component based on the registered first projection data and the second projection data, the scatter component including low-frequency scattered radiation signals.

Abstract: A method and apparatus is provided to reconstruct a computed tomography image using iterative reconstruction (IR) that is accelerated using various combinations of ordered subsets, conjugate gradient, preconditioning, resetting/restarting, and/or gradient approximation techniques. For example, when restarting criteria are satisfied the IR algorithm can be reset by setting conjugate-gradient parameters to initial values and/or by changing the number of ordered subsets. The IR algorithm can be accelerated by approximately calculating the gradients, by using a diagonal or Fourier preconditioner, and by selectively updating the preconditioner based on the regularization function. The update direction and step size can be calculated using the preconditioner and a surrogate function, which is not necessarily separable.

Abstract: An output device with which an operator can understand a progress state of machine learning from evaluation function values is provided. The output device includes: an information acquisition unit that acquires a plurality of evaluation function values which use servo data or are calculated using the servo data from a machine learning device that performs machine learning with respect to a servo control device that controls a servo motor that drives a shaft of a machine tool, a robot, or an industrial machine; and an output unit that outputs the plurality of acquired evaluation function values. The output unit may include a display unit that displays the plurality of evaluation function values on a display screen.

Abstract: Side coordinate groups that correspond to the four sides of the well are extracted from the edge coordinate group. Approximate lines are generated from the side coordinate groups. Corner coordinate groups that correspond to the four corners of the well are extracted from the edge coordinate group. Two intersecting approximate lines are selected from the approximate lines, and a contact candidate coordinates is set on each of the two approximate lines. Every possible pair of one contact candidate coordinate on one of the two approximate lines and one contact candidate coordinates set on the other approximate line are generated. For the generated pairs, processing for calculating an approximate curve by polynomial approximation from the corner coordinate group is performed, the approximate curve being in contact with two contact candidate coordinates. An approximate curve that is closest to the corner coordinate group is selected from all the calculated approximate curves.

Abstract: A method includes accessing a digital image in a raster representation and identifying edges in the digital image. A set of curves are generated by fitting each of the edges with a corresponding curve. A curvilinear triangulation is generated to include the set of curves. The method further includes generating a gradient mesh based on the curvilinear triangulation, where the gradient mesh includes polygons of the curvilinear triangulation along with control points based on the curvilinear triangulation. A respective color is determined for each control point of the gradient mesh, based on pixel colors of pixels of the raster representation corresponding to the control points in the gradient mesh. The method further includes generating a vector representation of the digital image, where the vector representation includes the gradient mesh with colors of the control points.

Abstract: A highly visible overlay system may include a display system and contrasting visible element configuration components configured to define contrasting visible elements. An imaging component is configured to capture an image of a target scene for display. A processing component is configured to construct the plurality of contrasting visual elements in accordance with a visual acuity factor, and generate an electronic overlay constructed of the contrasting visual elements. A display component is configured to display the constructed overlay in combination with the image of the target scene, with the overlay being displayed as an overlay on the image of the target scene. The contrasting visual elements may include a white block and a black block, configured to satisfy visual acuity factor, display configuration and field of view information.

Abstract: Systems and methods are provided for receiving an image from a camera of a mobile device, analyzing the image to determine a subject of the image, segmenting the subject of the image to generate a mask indicating an area of the image comprising the subject of the image, applying a bokeh effect to a background region of the image to generate a processed background region, generating an output image comprising the subject of the image and the processed background region, and causing the generated output image to display on a display of the mobile device.

Abstract: Techniques for automatically synchronizing poses of objects in an image or between multiple images. An automatic pose synchronization functionality is provided by an image editor. The image editor identifies or enables a user to select objects (e.g., people) whose poses are to be synchronized and the image editor then performs processing to automatically synchronize the poses of the identified objects. For two objects whose poses are to be synchronized, a reference object is identified as one whose associated pose is to be used as a reference pose. A target object is identified as one whose associated pose is to be modified to match the reference pose of the reference object. An output image is generated by the image editor in which the position of a part of the target object is modified such that the pose associated with the target object matches the reference pose of the reference object.