Abstract: A method and system of automatically estimating a ball carrier in team sports.

Abstract: The present disclosure provides a folding seat structure and a folding scooter. The folding seat structure includes a backrest mechanism, an armrest mechanism and a folding mechanism. The armrest mechanism includes an armrest assembly, an armrest elevating assembly, an armrest rotating assembly and an armrest moving assembly. The seat assembly is arranged at a top end of the folding mechanism. A bottom of the folding mechanism is hinged to a vehicle frame and is capable of being switched between an unfolded position and a folded position relative to the vehicle frame.

Abstract: A resource allocation method includes determining target resources needed by each of a plurality of large models for task processing. Different ones of the plurality of large models are used to process different types of tasks. The target resources include at least computing power resources. The method further includes predicting, based on historical records, a number of candidate tasks and types of the candidate tasks to be processed in a future time segment, determining, based on the number and the types, a resource allocation strategy for the future time segment according to the target resources needed by each of the plurality of large models for task processing, and allocating, based on the resource allocation strategy, initial target resources for each of the plurality of large models in the future time segment, and allocating reserved resources. The reserved resources are reserved from total target resources.

Abstract: Embodiments are generally directed to a Conditional Kernel Prediction Network (CKPN) for image and video de-noising and other related image and video processing applications. Disclosed is an embodiment of a method for de-noising an image or video frame by a convolutional neural network implemented on a compute engine, the image including a plurality of pixels, the method comprising: for each of the plurality of pixels of the image, generating a convolutional kernel having a plurality of kernel weights for the pixel, the plurality of kernel weights respectively corresponding to pixels within a region surrounding the pixel; adjusting the plurality of kernel weights of the convolutional kernel for the pixel based on convolutional kernels generated respectively for the corresponding pixels within the region surrounding the pixel; and filtering the pixel with the adjusted plurality of kernel weights and pixel values of the corresponding pixels within the region surrounding the pixel to obtain a de-noised pixel.

Abstract: Methods, articles, and systems herein use neural network-based reduced bit-depth image data input for video coding.

Abstract: The present disclosure provides a pharmaceutical formulation for oral administration comprising a solid dispersion comprising amorphous X842 and a water-soluble, amphiphilic carrier.

Abstract: A decoding method, a device, a storage medium and an application program are provided.

Abstract: Techniques related to 3D pose estimation from a 2D input image are discussed. Such techniques include incrementally adjusting an initial 3D pose generated by applying a lifting network to a detected 2D pose in the 2D input image by projecting each current 3D pose estimate to a 2D pose projection, applying a residual regressor to features based on the 2D pose projection and the detected 2D pose, and combining a 3D pose increment from the residual regressor to the current 3D pose estimate.

Abstract: The disclosure provides a communication delay compensation method and a communication delay compensation system based on an autonomous robot, where the method includes the following steps: establishing a state equation based on a system model of an the AUV positioning system; acquiring an included angle between a direction vector of the AUV to an observation station and a velocity vector of the AUV based on the system model; establishing an observation equation according to the state equation and the included angle; establishing an extended Kalman filter equation based on the system model, the included angle and the observation equation; and calculating a position information predicted value at the current time by using the extended Kalman filter equation to complete communication delay compensation.

Abstract: A battery cell, a battery, and an electrical apparatus are disclosed. The battery cell includes a shell having a first wall and a second wall joined together. The first wall is formed with a groove that ruptures during pressure relief to release internal pressure. The second wall is located adjacent to the first wall in a first direction, and the groove includes a segment aligned with the second wall in that direction. The first wall also has a second groove, and in the thickness direction of the first wall, the projection of the second groove lies between the projection of the groove segment and the outer surface of the second wall. This structure allows the second groove to absorb excess material produced during formation of the first groove segment, thereby reducing arching or loss of flatness in the first wall and improving stability of the battery cell.

Abstract: A method and system of image processing for action classification uses fine-grained motion-attributes.

Abstract: An approach for providing a privacy mask for video streams. The approach involves, for example, processing one or more images of a video stream to detect at least one object. The approach also determines one or more feature vectors from a plurality of pixels of the one or more images that correspond to the at least one detected object. The approach further involves encoding the one or more feature vectors into at least one mask that obscures the at least one object in the one or more images. The at least one mask is trackable across the one or more images of the video stream based on the one or more feature vectors. The approach further involves causing an addition of the at least one mask to the one or more images of the video stream and providing the video stream with the at least one mask as an output.

Abstract: The present disclosure provides an apparatus and method of guided neural network model for image processing. An apparatus may comprise a guidance map generator, a synthesis network and an accelerator. The guidance map generator may receive a first image as a content image and a second image as a style image, and generate a first plurality of guidance maps and a second plurality of guidance maps, respectively from the first image and the second image. The synthesis network may synthesize the first plurality of guidance maps and the second plurality of guidance maps to determine guidance information. The accelerator may generate an output image by applying the style of the second image to the first image based on the guidance information.

Abstract: The present disclosure provides an apparatus and method of guided neural network model for image processing. An apparatus may comprise a guidance map generator, a synthesis network and an accelerator. The guidance map generator may receive a first image as a content image and a second image as a style image, and generate a first plurality of guidance maps and a second plurality of guidance maps, respectively from the first image and the second image. The synthesis network may synthesize the first plurality of guidance maps and the second plurality of guidance maps to determine guidance information. The accelerator may generate an output image by applying the style of the second image to the first image based on the guidance information.

Abstract: A battery cell, a battery, and an electrical apparatus. The battery cell includes a shell and a pressure relief component. The shell includes a first wall. The pressure relief component is disposed on the first wall and provided with a first groove defining at least one predetermined pressure relief region, and the pressure relief component is configured to be capable of rupturing along at least part of the first groove during pressure relief of the battery cell. The pressure relief component is further provided with a second groove which is arranged in first direction perpendicular to a thickness direction of the first wall with the first groove, located on one side of the first groove in the first direction, and configured to guide at least part of the predetermined pressure relief region to flip over to open at least part of the predetermined pressure relief region.

Abstract: The present application discloses a method for early warning brandish of a transmission wire based on an improved Bayes-Adaboost algorithm, including: forming a classifier by training a historical brandish fault training set, and by using an Adaboost ensemble learning method, and obtaining an early warning result of the brandish of the transmission wire via the classifier according to real-time forecast meteorological information and information of different parameters of the transmission wire. The present invention can realize calculation and processing of forecast information of meteorological characteristic factors of the brandish of the transmission wire, structural parameters of the transmission wire and other related data, and obtain an early warning analysis result of a brandish disaster of the transmission wire in a region.

Abstract: Techniques related to 3D pose estimation from a 2D input image are discussed. Such techniques include incrementally adjusting an initial 3D pose generated by applying a lifting network to a detected 2D pose in the 2D input image by projecting each current 3D pose estimate to a 2D pose projection, applying a residual regressor to features based on the 2D pose projection and the detected 2D pose, and combining a 3D pose increment from the residual regressor to the current 3D pose estimate.

Abstract: Apparatus and methods for three-dimensional pose estimation are disclosed herein. An example apparatus includes an image synchronizer to synchronize a first image generated by a first image capture device and a second image generated by a second image capture device, the first image and the second image including a subject; a two-dimensional pose detector to predict first positions of keypoints of the subject based on the first image and by executing a first neural network model to generate first two-dimensional data and predict second positions of the keypoints based on the second image and by executing the first neural network model to generate second two-dimensional data; and a three-dimensional pose calculator to generate a three-dimensional graphical model representing a pose of the subject in the first image and the second image based on the first two-dimensional data, the second two-dimensional data, and by executing a second neural network model.

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed for real-time interactive performances.

Abstract: A scheduling method is applied to cloud services. The scheduling includes determining at least one target node in a target cluster, the target cluster including at least two nodes, each node including a host machine and at least one virtual machine connected to the host machine; determining a scheduling strategy for the virtual machine in the at least one target node, the scheduling strategy being used to schedule the virtual machine connected to the host machine in the target node; and reducing a frequency of the host machine in the target node based on the scheduling strategy.