Abstract: Embodiments of this application disclose a foreground data generation method performed at a computer device. The method includes: obtaining a background image and a target image, the target image containing a target object and a background; removing the background from the target image according to the background image and the target image, to obtain initial foreground data of the target object in the target image; obtaining certain foreground data and uncertain data from the initial foreground data, wherein the uncertain data represents data whose value is between the certain foreground data and background data corresponding to the background; and segmenting the certain foreground data from the uncertain data, to obtain target foreground data of the target object in the target image.

Abstract: First and second pluralities of residual elements useable to reconstruct first and second respective parts of a representation of a signal are obtained. A transformation operation is performed to generate at least one correlation element. The transformation operation involves at least one residual element in the first plurality and at least one residual element in the second plurality. The at least one correlation element is dependent on an extent of correlation between the at least one residual element in the first plurality and the at least one residual element in the second plurality. The transformation operation is performed prior to the at least one correlation element being encoded.

Abstract: Sensors, including time-of-flight sensors, may be used to detect objects in an environment. In an example, a vehicle may include a time-of-flight sensor that images objects around the vehicle, e.g., so the vehicle can navigate relative to the objects. Sensor data generated by the time-of-flight sensor can include returns associated with highly reflective objects that cause glare. In some examples, a depth of a sensed surface is determined from the sensor data and additional pixels at the same depth are identified. The subset of pixels at the depth are filtered by comparing a measured intensity value to a threshold intensity value for the depth. Other threshold intensity values can be applied to subsets of pixels at different depths.

Abstract: An object detection system includes at least one memory storing instructions, and at least one processor that, upon executing instructions stored in the memory, controls the object detection system to perform functions including inputting a first video frame and a second video frame from a camera; generating a first set of predicted object locations using only one of the first video frame and the second video frame; generating a second set of predicted object locations based on pixel differences between the first video frame and the second video frame; and determining a final set of object locations based on the first set of predicted object locations and the second set of predicted object locations.

Abstract: An embodiment of the present invention is directed to a combination of two deep-learning computer vision models—customized with post-processing—wrapped in a mobile application that is backed by an Application Programming Interface (API) supporting concurrent mobile users to accomplish asset serialization tasks in a warehouse or other storage environment.

Abstract: A face recognition method based on an evolutionary convolutional neural network is provided. The method optimizes the design of convolutional neural network architecture and the initialization of connection weights by using a genetic algorithm and finds an optimal neural network through continuous evolutionary calculation, thus reducing dependence on artificial experience during the design of the convolutional neural network architecture. The method encodes the convolutional neural networks by using a variable-length genetic encoding algorithm, so as to improve the diversity of structures of convolutional neural networks. Additionally, in order to cross over extended chromosomes, structural units at corresponding positions are separately crossed over and then recombined, thereby realizing the crossover of chromosomes with different lengths.

Abstract: Methods for operating a motion recognition apparatus are disclosed. In some implementations, a method for recognizing a motion or gesture of an object may include operating an optical sensor device to capture light reflected from the object under illumination by light emitted toward the object, generating, by comparing the emitted light to the reflected light, a depth image including distance information indicating a distance between the optical sensor device and the object, generating, based on the light reflected from the object, an infrared image including infrared image information associated with the object, and determining the motion of the object based on at least one of the depth image and the infrared image.

Abstract: An approach is provided in which the approach receives an image that includes multiple image points and constructs a plane in the image based on a first subset of the plurality of image points. The approach identifies a second subset of the image points that belong to the plane and are not part of the first subset of image points, and removes the first subset of image points and the second subset of image points form the image points. The approach annotates the remaining subset of image points in the image.

Abstract: A G-PCC encoder and G-PCC decoder may quantize and scale, respectively, a position of a child node. The G-PCC encoder may control the precision of the quantization and scaling using a quantization parameter (QP) value and a parameter value k, wherein the parameter value k specifies a number of QP points per doubling of a scaling step size to be used at the G-PCC decoder.

Abstract: A detection device includes: a first detector which irradiates light onto the target object and detects light emitted from the target object; a first shape information generator which generates first shape information representing a first shape of the target object on the basis of a detection result of the first detector; and a second shape information generator which adds a second shape, which is based on information different from the detection result of the first detector, to the first shape, and which generates second shape information representing a shape including the first shape and the second shape.

Abstract: Techniques related to accelerated video enhancement using deep learning selectively applied based on video codec information are discussed. Such techniques include applying a deep learning video enhancement network selectively to decoded non-skip blocks that are in low quantization parameter frames, bypassing the deep learning network for decoded skip blocks in low quantization parameter frames, and applying non-deep learning video enhancement to high quantization parameter frames.

Abstract: A system and method for identifying an object to be picked up by a robot. The method includes obtaining a 2D red-green-blue (RGB) color image and a 2D depth map image of the objects using a 3D camera, where pixels in the depth map image are assigned a value identifying the distance from the camera to the objects. The method generates a segmentation image of the objects using a deep learning convolutional neural network that performs an image segmentation process that extracts features from the RGB image, assigns a label to the pixels so that objects in the segmentation image have the same label and rotates the object using the orientation of the object in the segmented image. The method then identifies a location for picking up the object using the segmentation image and the depth map image and rotates the object when it is picked up.

Abstract: An image is acquired by a camera. The image has a first set of characters and a second set of characters. The first set of characters are classified as an identifier. The second set of characters are classified as data associated with the identifier. The image is divided to create an image segment. The image segment includes the first set of characters and not the second set of characters. The first set of characters are decoded in the image segment to generate a first character string. The second set of characters are decoded to generate a second character string. The first character string is linked to the second character string based on classifying the first set of characters as the identifier and the second set of characters as the data associated with the identifier.

Abstract: A vehicle information detection method, a method for training a detection model, an electronic device and a storage medium are provided, and relates to the technical field of artificial intelligence, in particular to the technical field of computer vision and deep learning. The method includes: performing a first target detection operation based on an image of a target vehicle, to obtain a first detection result for target information of the target vehicle; performing an error detection operation based on the first detection result, to obtain error information; and performing a second target detection operation based on the first detection result and the error information, to obtain a second detection result for the target information.

Abstract: Sensors, including time-of-flight sensors, may be used to detect objects in an environment. In an example, a vehicle may include a time-of-flight sensor that images objects around the vehicle, e.g., so the vehicle can navigate relative to the objects. The sensor may generate first image data at a first configuration and second image data at a second configuration. A disambiguated depth of a surface may be determined from the first image data and the second image data. If the disambiguated depth is greater than a nominal maximum depth of the sensor in the first configuration, an intensity of the surface may be determined from the first image data. If the intensity meets or exceeds a threshold intensity, the surface may be determined to be beyond the nominal maximum depth. If the intensity is less than the threshold intensity, an actual depth of the surface may be determined form the second image data as a distance less than the nominal maximum depth.

Abstract: Systems and methods for image segmentation using neural networks and active contour methods in accordance with embodiments of the invention are illustrated. One embodiment includes a method for generating image segmentations from an input image. The method includes steps for receiving an input image, identifying a set of one or more parameter maps from the input image, identifying an initialization map from the input image, and generating an image segmentation based on the set of parameter maps and the initialization map.

Abstract: The invention provides an offline handwriting individual recognition system and method based on two-dimensional dynamic features.

Abstract: Systems and methods for image processing for determining a registration map between a first image of a scene with a second image of the scene, include solving an optimal transport (OT) problem to produce the registration map by optimizing a cost function that determines a minimum of a ground cost distance between the first and the second images modified with an epipolar geometry-based regularizer including a distance that quantifies the violation of an epipolar geometry constraint between corresponding points defined by the registration map. The ground cost compares a ground cost distance of features extracted within the first image with a ground cost distance of features extracted from the second image.

Abstract: A method comprising obtaining, from a sensor, depth data representing a target object; selecting a model to fit to the depth data; for each data point in the depth data: defining a ray from a location of the sensor to the data point; and determining an error based on a distance from the data point to the model along the ray; when the depth data does not meet a similarity threshold for the model based on the determined errors, selecting a new model and repeating the error determination for the depth data based on the new model; when the depth data meets the similarity threshold for the model, selecting the model as representing the target object; and outputting the selected model representing the target object.

Abstract: Provided herein are MD Codes, MD DYNA Codes, MD ASA, and Next Human system, and the methods of using them to diagnose and treat aesthetic conditions or disorders.