Abstract: An edge detecting device includes a feature extracting circuit configured to extract first and second feature data from an input image; a prototype generating circuit configured to generate prototype data using the first feature data and an input label, the prototype data including foreground and background information of an object; a region detecting circuit configured to generate a segmentation mask by detecting a region of an object using the first feature data and the prototype data; and an edge extracting circuit configured to generate an edge map by combining the segmentation mask and the second feature data.

Abstract: A method for computing, a computing device, and a computer-readable storage medium are provided. The method includes determining a first pixel block in a cache. The first pixel block is composed of a 2m row×2n column pixel matrix and includes original pixel data and pixel data related to the original pixel data. The first pixel block is read from the cache. At least part of the pixel data related to the original pixel data is used for padding related to the original pixel data. The original pixel data includes pixel data from the (n+1)th column to the 2nth column in the (m+1)th row to the 2mth row in the 2m row×2 n column pixel matrix. When reading data from the cache, pixel data that needs to be obtained after insert-zero and padding operations on the original pixel data in back propagation can be read at one time.

Abstract: In the athletic competition involving jumping such as figure skating, it is desired to measure jumping height and flight distance by calculating a trajectory of a moving object as a target from moving images captured by a monocular camera. Provided is a device, a method and a program to calculate the trajectory of the moving object jumping in three dimensions from information of a plurality of image frames captured by the monocular camera by detecting a specific point of the moving object, calculating an amount of change with respect to three-dimensional positions of the specific point in the consecutive image frames, and calculating the trajectory of the specific point from a positional relation between straight lines having the positions of the specific point and a curved line capable of expressing a parabolic motion passing through a take-off point and a landing point.

Abstract: A pure pose solution method and system for a multi-view camera pose and scene are provided. The method includes: a pure rotation recognition (PRR) step: performing PRR on all views, and marking views having a pure rotation abnormality, to obtain marked views and non-marked views; a global translation linear (GTL) calculation step: selecting one of the non-marked views as a reference view, constructing a constraint tr=0, constructing a GTL constraint, solving a global translation (I), reconstructing a global translation of the marked views according to tr and (I), and screening out a correct solution of the global translation; and a structure analytical reconstruction (SAR) step: performing analytical reconstruction on coordinates of all 3D points according to a correct solution of a global pose. The method and system can greatly improve the computational efficiency and robustness of the multi-view camera pose and scene structure reconstruction.

Abstract: A detection method for an image segmentation manner is provided. The detection method includes obtaining four sub-images input by a serial digital interface and formed by dividing an original image, extracting four groups of reference pixel points from the four sub-images according to a set extraction manner, calculating pixel difference values of pixel points according to the four groups of reference pixel points, and determining an image segmentation manner of the serial digital interface according to the pixel difference values.

Abstract: This document describes systems and techniques for clustering track pairs for multi-sensor track association. Many track-association algorithms use pattern-matching processes that can be computationally complex. Clustering tracks derived from different sensors present on a vehicle may reduce the computational complexity by reducing the pattern-matching problem into groups of subproblems. The weakest connection between two sets of tracks is identified based on both the perspective from each track derived from a first sensor and the perspective of each track derived from a second sensor. By identifying and pruning the weakest connection between two sets of tracks, a large cluster of tracks may be split into smaller clusters. The smaller clusters may require fewer computations by limiting the quantity of candidate track pairs to be evaluated. Fewer computations result in processing the sensor information more efficiently that, in turn, may increase the safety and reliability of an automobile.

Abstract: The disclosed computer-implemented method may include (1) accessing a video portraying an object within a set of frames, (2) defining a subset of key frames within the video based on movement of the object across the set of frames, (3) generating, for each key frame within the subset of key frames, a spline outlining the object within the key frame, (4) receiving input to adjust, for a selected key frame within the subset of key frames, a corresponding spline, and (5) interpolating the adjusted spline with a spline in a sequentially proximate key frame to define the object in frames between the selected key frame and the sequentially proximate key frame. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Provided is an image compression method for compressing image data generated by an image sensor, the image compression method including detecting a saturated pixel among a plurality of pixels included in a pixel group included in the image data, the saturated pixel having a pixel value exceeding a threshold value, and the plurality of pixels being adjacent to each other and having a same color as each other, generating a saturation flag indicating a position of the saturated pixel, compressing the image data by comparing a reference pixel with at least one non-saturated pixel among the plurality of pixels included in the pixel group, and outputting a bitstream including the saturation flag, a compression result, and a compression method.

Abstract: Described herein are a system and methods for generating a record of objects, as well as respective positions for those objects, with respect to a user. In some embodiments, a user may use a user device to scan an area that includes one or more objects. The one or more objects may be identified from image information obtained from the user device. Positional information for each of the one or more objects may be determined from depth information obtained from a depth sensor installed upon the user device. In some embodiments, the one or more objects may be mapped to object models stored in an object model database. The image information displayed on the user device may be augmented so that it depicts the object models associated with the one or more objects instead of the actual objects.

Abstract: The application provides an image acquisition apparatus, a server, and encryption and decryption methods, and relates to the field of data processing. The image acquisition apparatus includes: an image acquisition device including an encryption processor and a data interface provided in the image acquisition device, wherein the encryption processor is configured to encrypt, by using a stored first key, image data acquired by the image acquisition device to obtain image ciphertext data; the data interface is configured to obtain the image ciphertext data and output the image ciphertext data to a terminal device. With the technical solution of the application, security of the image data can be ensured.

Abstract: Embodiments of the invention provide a laser system including a laser station including at least one laser, at least one processor, and at least one non-transitory computer-readable storage medium in data communication with the at least one processor. The at least one non-transitory computer-readable storage medium includes program instructions executable by the at least one processor, and enabling or operating an exchange of data between the at least one laser station and at least one remote network. The laser system includes at least one GUI display configured and arranged to display operating parameters or functions of the at least one laser and/or any of the data exchanged between the laser station and at least one remote network. The program instructions include instructions that direct the processor to update the at least one GUI display with a plurality of user-selectable graphics arranged around the periphery or circumference of a central display.

Abstract: A method of constructing a transformer model for answering questions about a video story according to an embodiment includes: extracting feature vectors related to each character of a video from video data including vision data and subtitle data and question data for video questions and answers, and generating an input embedding using the feature vectors related to the character; and training a transformer model using the input embedding.

Abstract: A structural analysis computing device may generate a proposed insurance claim and/or generate a proposed insurance quote for an object pictured in a three-dimensional (3D) image. The structural analysis computing device may be coupled to a drone configured to capture exterior images of the object. The structural analysis computing device may include a memory, a user interface, an object sensor configured to capture the 3D image, and a processor in communication with the memory and the object sensor. The processor may access the 3D image including the object, and analyze the 3D images to identify features of the object—such as by inputting the 3D image into a trained machine learning or pattern recognition program. The processor may generate a proposed claim form for a damaged object and/or a proposed quote for an uninsured object, and display the form to a user for their review and/or approval.

Abstract: Described herein is a system and computer implemented method for finding similars for a selected clothing image amongst a set of clothing images in an electronic catalog in an online store serving online customers. An object detection model is applied to extract the clothing section within the clothing images to create a preprocessed image. A first machine learning model is applied on the preprocessed image(s) to convert the colors and textures of said preprocessed image into a first set of vector representations. A second machine learning model is applied on the preprocessed image(s) to convert the shapes of said preprocessed image into a second set of vector representations. Operations of mapping nearest vectors, matching attributes, sorting and ranking are performed, and thereafter similar images are displayed to the online customer.

Abstract: The present application provides a light sensor and a display device. A light sensing transistor and a switching transistor in the light sensor are configured to form a light sensor circuit. In a structural design, amorphous silicon is configured as a first active pattern of the light sensing transistor, so that a thickness of a channel region of the first active pattern is greater than or equal to 5000 angstroms. Therefore, a number of photo-generated carriers of the light sensing transistor is increased, which makes the light sensor have higher responses and increases fingerprint or palmprint recognition success rates.

Abstract: In some implementations, a method includes: obtaining a first image; determining characterization properties for each of a plurality of pixels within the first image; determining a dominant hue based on the one or more characterization properties for each of the plurality of pixels within the first image; determining a plurality of hues that satisfy a predetermined perception threshold relative to the dominant hue based on the characterization properties for each of the plurality of pixels within the first image, wherein the plurality of hues is different from the dominant hue; and generating a second image based at least in part on the dominant hue and the plurality of hues. In some implementations, the second image corresponds to a color gradient generated based on the dominant hue within the first image. In some implementations, the second image is set as a wallpaper for a user interface.

Abstract: An image-based real-time intrusion detection method and a surveillance camera that use artificial intelligence are provided to sample a plurality of frames input at a first point in time, acquire a probability that at least one object corresponding to a type of a target object exists in an image of the respective sampled frames by using a first artificial neural network, adjust a sampling rate for a plurality of frames to be input at a second point in time after the first point in time according to processing time of each frame of the first artificial neural network required to acquire an existence probability of the at least one object, select each of the respective sampled frames as a frame of the target object according to a magnitude of the acquired probability, generate a movement trajectory of each object corresponding to the type of the target object from the frames selected as the frame of the target object, and acquire an intrusion occurrence probability from the generated movement trajectory by usin

Abstract: A method for processing images of animals obtains first training data, which includes first images of animals and first annotation data including first location data for predefined feature points on at least one animal in each image in the first images, and trains a first machine learning-based model (MLM) by use of the first training data to generate a first trained MLM. The method obtains second images of animals, operates the first trained MLM on the second images to generate second annotation data comprising second location data for the predefined feature points on at least one animal in each image in the second images, and aggregates at least a subset of the first and second images, and a corresponding subset of the first and second annotation data, into second training data, which thereby is expanded in relation to the first training data.

Abstract: In order to determine whether a color contact lens is being worn on an iris, this image processing device includes: a first feature amount calculation means for using an image of the iris in a first wavelength band to calculate a first feature amount in the first wavelength band, said iris being used in authentication of a person; a second feature amount calculation means for using an image of the iris in a second wavelength band to calculate a second feature amount in the second wavelength band; and a covering determination means for determining, on the basis of the first feature amount and the second feature amount, whether a color contact lens is present on the iris.

Abstract: Methods for characterizing fluids from a patient. A time series of images of a conduit are received, and a conduit image region in the images is identified. A flow type of the fluids passing through the conduit may be classified as one of air, laminar liquid, and turbulent liquid by evaluating an air-liquid boundary of the fluid. A volumetric flow rate of the fluids in the conduit is estimated. The volumetric flow rate may be based on the classified flow type. A concentration of a blood component of the fluids passing through the conduit may be estimated based on the images. A proportion of the fluid that is blood may also be determined, and a volume of blood that has passed through the conduit within a predetermined period of time may be estimated based on the estimated total volumetric flow rate and the determined proportion.