Motion vector prediction method, video coding method, and related equipment and devices

Abstract

The present application discloses a motion vector prediction method, a video encoding method, and related equipment and devices, wherein the motion vector prediction method includes: obtaining a candidate list of the current coding block in a preset prediction mode, wherein the candidate list includes multiple candidate motion vectors; perform motion search on multiple candidate motion vectors in the candidate list to obtain the initial optimal motion vector of the current coding block in the preset prediction mode; use the first preset number of pixel precisions to respectively perform initial optimal motion vectors Perform motion compensation on the vector to determine the final optimal motion vector of the current coding block; wherein, the preset prediction mode includes any one of multiple prediction modes, and the pixel precision used in any one of the prediction modes is the same. The above solution can reduce the difficulty of realizing the motion vector prediction by software and hardware, thereby improving the prediction accuracy.

Description

technical field [0001] The present application relates to the field of video coding and decoding, and in particular to a motion vector prediction method, a video coding method, and related equipment and devices. Background technique [0002] Due to the relatively large amount of video image data, it is usually necessary to encode and compress video pixel data (such as RGB, YUV, etc.) to reduce the amount of video data. The compressed data is called a video code stream, and the video code stream passes through a wired or wireless network. It is transmitted to the client, and then decoded and viewed. [0003] Video coding mainly includes acquisition, prediction, transform quantization and entropy coding. Among them, the prediction part is divided into three parts: intra prediction, inter prediction and intra block copy prediction. However, there are differences between different prediction modes, which brings difficulties to the realization of prediction software and hardwar...

AI Technical Summary

Problems solved by technology

However, there are differences between different prediction modes, which brings difficulties to the realization of prediction software and hardware.

For example, as one of the main prediction modes, AMVP (Advanced Motion Vector Prediction, Advanced Motion Vector Prediction) mode has developed multiple branches in the process of technological development, for example, conventional AMVP mode, affine AMVP mode, frame-based Inner block copy AMVP mode, hash AMVP mode, etc., and in the process of selecting and using from the multiple modes included in the above-mentioned AMVP mode, especially when more than one mode is selected, due to various The difference between modes requires a separate processing loop for each mode, which is not conducive to the realization of software and hardware, and increases the difficulty of implementation

Images