Techniques for training and using machine learning models for video encoding mode selection are described. According to some embodiments, a computer-implemented method includes receiving a live video at a content delivery service, extracting one or more features for a plurality of macroblocks of a frame of the live video, determining an encoding mode from a plurality of encoding modes for each of the plurality of macroblocks of the frame with a machine learning model based at least in part on an input of the one or more features, performing a real time encode of the frame of the live video based at least in part on the determined encoding modes to generate an encoded frame by the content delivery service, and transmitting the encoded frame from the content delivery service to a viewer device.

Disclosed herein includes a system, a method, and a device for reducing latency and jitter by reserving a transmission duration for a transmit opportunity. A device can compress a first slice of a plurality of slices of a video frame to obtain a size of the first slice after compression. The device can estimate a size of the video frame after compression, according to the size of the first slice after compression. The device can generate a request to send (RTS) to reserve, a transmission channel for a transmission duration corresponding to the size of the video frame after compression. The device can broadcast the RTS to one or more other devices.

In some aspects, the disclosure is directed to methods and systems for reducing memory utilization and increasing efficiency during affine merge mode for versatile video coding by utilizing motion vectors stored in a motion data line buffer for a prediction unit of a second coding tree unit neighboring a first coding tree unit to derive control point motion vectors for the first coding tree unit.

The entropy coding of a current part of a predetermined entropy slice is based on, not only, the respective probability estimations of the predetermined entropy slice as adapted using the previously coded part of the predetermined entropy slice, but also probability estimations as used in the entropy coding of a spatially neighboring, in entropy slice order preceding entropy slice at a neighboring part thereof. Thereby, the probability estimations used in entropy coding are adapted to the actual symbol statistics more closely, thereby lowering the coding efficiency decrease normally caused by lower-delay concepts. Temporal interrelationships are exploited additionally or alternatively.

An image evaluation device includes a determination result acquisition unit acquires a result of determining the presence or absence of a difference between an object image that is one of a plurality of images that include three or more images obtained by imaging substantially the same spatial region and each of reference images that are images other than the object image among the plurality of images and an evaluation index acquisition unit configured to acquire an evaluation index for the plurality of images on the basis of at least one of the number of determinations of the presence of the difference between the object image and each reference image and the number of determinations of the absence of the difference between the object image and each reference image.

An image encoding/decoding method and apparatus are provided. An image decoding method performed by an image decoding apparatus may comprise parsing weight information specifying a weight for a reference sample from a bitstream according to a weight parameter syntax structure, and decoding a current block by performing inter prediction based on the weight information. The parsing according to the weight parameter syntax structure may comprise obtaining weight number information specifying the number of weight information obtained from the bitstream according to the weight parameter syntax structure and obtaining weight information from the weight parameter syntax structure based on the weight number information.

Methods, systems, and devices for luma mapping with chroma scaling for video and image coding are disclosed. An example method of video processing includes performing, for a current region comprising a luma block, a first chroma block, and a second chroma block, a conversion between the current region of a video and a bitstream representation of the video according to a rule that specifies an order in which, during decoding, the first chroma block and the second chroma block are processed based on mapped sample values of the luma block.

A method comprising obtaining a bitstream, the bitstream comprises a transform unit syntax and a coding unit syntax, the transform unit syntax includes a value of a first flag and a value of a second flag related to, respectively, a first chroma transform block and a second chroma transform block of a current transform unit or a current sub-transform unit within the current transform unit, the first or second flag specifies whether the first or second chroma transform block contains at least one transform coefficient levels not equal to 0, the coding unit syntax includes a value of a third flag specifying whether a transform tree structure is present or not; and deriving a value of a fourth flag based on the values of the first, second, and third flags, the fourth flag specifies whether a luma transform block contains at least one transform coefficient levels not equal to 0.

This application discloses a motion vector obtaining method and apparatus, a computer device, and a storage medium. In the method, an initial motion vector of a to-be-processed picture block is determined by using a location relationship between a reference block and the to-be-processed picture block. When the reference block and the to-be-processed picture block are located in a same coding tree block, a decoder uses an initial motion vector of the reference block as the initial motion vector of the to-be-processed picture block. When the reference block and the to-be-processed picture block are located in different coding tree blocks, the decoder uses a final motion vector of the reference block as the initial motion vector of the to-be-processed picture block.

An image decoding method according to the present invention includes reconstructing a residual block by inverse-quantizing and inverse-transforming an entropy-decoded residual block, generating a prediction block by performing intra prediction on a current block, and reconstructing an picture by adding the reconstructed residual block to the prediction block, wherein generating the prediction block includes generating a final prediction value of a prediction target pixel included in the current block based on a first prediction value of the prediction target pixel and a final correction value calculated by performing an arithmetic right shift on a two's complementary integer representation for an initial correction value of the prediction target pixel by a binary digit of 1. Accordingly, upon image encoding/decoding, computation complexity may be reduced.