Abstract: The application relates to refined quantization steps in video coding. A method for video processing includes: determining, during a conversion between a current video block and a bitstream representation of a video including the current video block, two-level quantization parameters comprising a first level quantization parameter QP1 and a second level quantization parameter QP2 for the current video block; and performing, at least based on the two-level quantization parameters, a first level quantization and/or a second level quantization during the conversion, wherein the two-level quantization parameters correspond to different quantization precisions respectively.

Abstract: A system and method are provided for inter-ceding video in which encoder and decoder memory requirements associated with storage of motion information related to collocated coding units is reduced. In some embodiments motion information related to only a single collocated coding unit may be stored at the encoder and decoder. In operation, if the encoder determines that motion information for a current coding unit should replace the currently stored motion information for the currently stored motion information for the collocated coding unit, then the encoder can replace the motion information at the encoder and transmit an indicator with the current coding unit to signal to the decoder that the motion information currently stored should be updated or replaced with the motion information associated with the current coding unit.

Abstract: A method and system for more efficient federated learning (FL) of a machine learning (ML) model using user equipment (UEs) in cellular networks are disclosed. In particular, a system is provided for reducing the impact of poor channel conditions in a cellular network on the FL process. The cellular network may be a 5G, 6G or next generation cellular network. Advantageously, this disclosure creates redundancies in the transmission of FL trained model parameters to reduce the likelihood of an FL training process being stalled by a failure in transmission of data between UEs and a central parameter server which updates an ML model using data received from UEs.

Abstract: The present invention relates to the field of image processing. Disclosed are a neural network-based pose estimation and registration method for heterogeneous images. In the present invention, a phase correlation algorithm is optimized to be differentiable and embedded into an end-to-end learning network framework, and a neural network-based pose estimation method for heterogeneous images is constructed. According to the method, an optimal feature extractor can be found for a result of image matching, a solution can be obtained without exhaustive evaluation, and good interpretability and generalization capability are achieved. The test results show that the present invention allows for accurate pose estimation and registration for heterogeneous images and shortening of the required time, has high accuracy and real-time performance, can meet actual application requirements, and can be applied in fields such as robot self-positioning.

Abstract: A display panel and a display device are provided. The display panel includes: a pixel unit including a pixel circuit and a light-emitting element, the pixel circuit including a driving transistor and a data writing transistor; and a data line connected to the data writing transistor; the data line includes a plurality of first-type data lines and a plurality of second-type data lines, each of the plurality of second-type data lines includes a first portion, a second portion, and a third portion, and the first portion and the second portion are connected by the third portion, the third portion and the second portion are located in different layers, the third portion and the first portion are located in different layers, the first portion is closer to the base substrate than the third portion, and the second portion is closer to the base substrate than the third portion.

Abstract: The present disclosure provides an ultrasonic imaging method and apparatus. The apparatus includes an ultrasonic signal emission source and a receiving array including ultrasonic signal receiving circuits; the method includes: turning on the source to perform emission of ultrasonic waves toward an object to be detected and causing the ultrasonic waves to propagate through the object, a depth-wise direction of which is along a propagation direction of the ultrasonic waves; after a first predetermined time period having elapsed since the source was turned on, turning on the receiving array to receive reflected echoes returning from a first section plane of the object to be detected perpendicular to the depth-wise direction; thereafter, turning off the receiving array, storing the reflected echoes by the ultrasonic signal receiving circuits and acquiring reflected echo signals, and successively reading the reflected echo signals from the ultrasonic signal receiving circuits during a reading time period.

Abstract: Techniques for implementing video processing techniques are described. In one example implementation, a method of video processing includes determining, for a conversion between a current block of a video and a bitstream representation of the video, whether a combined inter and intra prediction (CIIP) coding technique is to be applied to the current block based on a characteristic of the current block. The CIIP coding technique uses an intermedia inter prediction value and an intermedia intra prediction value to derive a final prediction value of the current block. The method also includes performing the conversion based on the determining.

Abstract: An autonomous mobile device (AMD) may perform tasks within a physical space. The AMD may move over ramps, bumps, or navigate around obstacles. The AMD may have an inertial measurement unit (IMU) and distance sensors. The IMU provides tilt information indicative of the AMD being on a flat surface or a ramp. The distance sensors provide information on distances between the AMD and surrounding obstacles. Using IMU measurements, the AMD determines a first speed limit that is safe given the tilt of the AMD. Using the distance sensors, the AMD determines a second speed limit that is safe given a distance to an obstacle. The AMD determines a maximum speed based on the first and second speed limits. Based on the maximum speed, the AMD determines whether to adjust a current speed and by how much.

Abstract: Motion Vector Management for Decoder Side Motion Vector Refinement is disclosed. A method of video processing including: determining, for a conversion between a first block of video and a bitstream representation of the first block of video, whether and/or how to apply decoder-side motion vector refinement (DMVR) based on a signaled information; and performing the conversion based on the determination.

Abstract: The present disclosure provides a method and apparatus for acquiring a pre-trained model, an electronic device and a storage medium, and relates to the field of artificial intelligence, such as the natural language processing field, the deep learning field, or the like. The method may include: adding, in a process of training a pre-trained model using training sentences, a learning objective corresponding to syntactic information for a self-attention module in the pre-trained model; and training the pre-trained model according to the defined learning objective. The solution of the present disclosure may improve a performance of the pre-trained model, and reduce consumption of computing resources, or the like.

Abstract: Devices, systems and methods for digital video coding, which includes methods for transform design, are described. In a representative aspect, a method of video processing includes performing a conversion between a current video block of a video and a bitstream of the video based on a rule, wherein the rule specifies that a selection of a transform matrix set for performing a transform operation during the conversion is based on a low-frequency non-separable transform index indicated in the bitstream, wherein the rule specifies that the transform operation includes, during an encoding operation, coding the current video block into the bitstream by applying a forward transform on residual values of the current video block, or wherein the rule specifies that the transform operation includes, during a decoding operation, generating from the bitstream the current video block by applying an inverse transform to scaled coefficients indicated in the bitstream.

Abstract: This disclosure provides a network resource pre-allocation method, device, system, and medium, wherein the method includes: an NWDAF acquiring historical record information of a user accessing a service, and generating user service preference prediction and suggestion information according to the historical record information; and the NWDAF sending the user service preference prediction and suggestion information to a network function (NF) so that the NF pre-allocates a network resource according to the user service preference prediction and suggestion information.

Abstract: A method, computer system, and a computer program product is provided for dynamic allocation of resources. A database access pattern is determined by analyzing and monitoring traffic pattern between a pool of resources and a plurality of clients. The relationship between each of the resources is also determined. Access is enabled to a plurality of resources based on the database access and resource relationships, so that the plurality of resources can be accessed but not allocated until processing a request. A consumption model is generated that predicts resource need during a processing request based on the resource relationships, traffic pattern and resources availability. Upon receipt of a subsequent request for processing, consumption model is used to predict resource needs and to dynamically allocate, re-allocate and release the plurality of resources in a cascading manner until completion of subsequent processing request.

Abstract: An array substrate includes: a first substrate (10), including a plurality of sub-pixel regions (101) arranged in an array along a row direction (X) and a column direction (Y); a pixel circuit layer, including a plurality of sub-pixel circuits; a planarization layer (17), provided with a first via hole (170) located in the sub-pixel regions (101), and includes at least one pattern portion (171), the pattern portion (171) includes a plurality of pattern units (171a) arranged in an array along the row direction (X) and the column direction (Y); and a reflective electrode layer, wherein the reflective electrode layer includes a plurality of reflective electrodes (18) that are mutually disconnected, each of the reflective electrodes (18) is located in one of the sub-pixel regions (101) and is electrically connected to the sub-pixel circuit through the first via hole (170).

Abstract: Disclosed is a software PUF based on an RISC-V processor for IoT security. A 32-bit RISC-V processor is used to generate abnormal information results in an abnormal operating state under a low voltage, and the abnormal information results are used to represent the features of the 32-bit RISC-V processor; 5-bit binary data obtained by comparing the abnormal information results with normal information results has high randomness and uniqueness and it is extremely difficult to directly extract internal abnormal information result from a hardware circuit of the 32-bit RISC-V processor, so modeling attacks based on the 5-bit binary data calculated according to the abnormal information results of the 32-bit RISC-V processor are almost impossible; in addition, when the 32-bit RISC-V processor is in an abnormal operating state, the operating frequency of the 32-bit RISC-V processor is dynamically adjusted through a frequency compensation method.

Abstract: A computer-implemented method according to one embodiment includes identifying a job creation request within a system; determining a current amount of available resources within the system; and conditionally deploying a reusable group of containers for the job, based on the current amount of available resources for the system.

Abstract: Provided is a display panel. The display panel includes: a substrate, at least one row of light-emitting elements disposed on the substrate, a plurality of pixel drive circuits, a plurality of connection lines, and a plurality of compensation portions. A length of a first connection line electrically connected to a first pixel drive circuit is less than a length of a second connection line electrically connected to a second pixel drive circuit. A capacitance of the first connection line is compensated by electrically connecting the compensation portion to the first connection line, and thus a difference between the capacitance of the first connection line and a capacitance of the second connection line is reduced.

Abstract: A method of visual media processing includes determining, for a conversion between a current video block of visual media data and a bitstream representation of the current video block, a buffer that stores reference samples for prediction in an intra block copy mode; for a sample spatially located at location of the current video block relative to an upper-left position of a coding tree unit including the current video block and having a block vector, computing a corresponding reference in the buffer at a reference location, wherein the reference location is determined using the block vector and the location; and upon determining that the reference location lies outside the buffer, re-computing the reference location based at least in part on a location of the current video block relative to the coding tree unit including the current video block.

Abstract: Methods, systems, and devices for coefficient scaling for high-precision image and video coding are described. An example method of video processing includes performing a conversion between a current block of a video and a bitstream representation of the video according to a rule. The rule specifies that the conversion includes during encoding, skipping applying a forward transform to residual coefficients of the current block prior to including in the bitstream representation, or during decoding, reconstructing residual coefficients of the current block from the bitstream representation without applying an inverse transform. The rule further specifies that a scale factor is applied to the residual coefficients independent of a size of the current block.

Abstract: History-based affine parameters inheritance is described. In an exemplary aspect, a method for video processing includes deriving, for a conversion between a current block of video and a bitstream representation of the current block, affine related information associated with an affine model to be used by the current block based on affine related information associated with an affine model used by a first block stored in a buffer, wherein the first block is affine coded prior to the current block; and performing the conversion by using the derived affine related information.