An Intelligent Affine Motion Compensation Coding Method

Abstract

The invention discloses an intelligent affine motion compensation encoding method, which includes: constructing an intelligent affine motion compensation network, the input of which is a space-time domain prediction block, a time domain reference block, and the output of motion field information is a current encoding block The prediction results; through the multi-scale convolution unit to extract features from the three inputs, design a motion estimation unit feature cascade, through two convolutional layers, respectively estimate the corresponding space-time domain prediction block and time-domain reference block Sports field information; using the sports field information, the feature map of the target image is obtained by compensating the characteristics of the space-time domain prediction block and the time domain reference block through a deformable convolution operation; taking the feature map of the target image as input, multi-channel information is fused based on attention and Reconstruct the target image. The present invention estimates a more accurate pixel-level sports field by making full use of the space domain, time domain and sports field information by virtue of the feature expression ability of deep learning.

Description

technical field [0001] The invention relates to the fields of deep learning and video coding, in particular to an intelligent affine motion compensation coding method. Background technique [0002] As high-definition and ultra-high-definition applications enter people's lives, the amount of video data continues to increase. The goal of video compression coding is to reduce the bit rate required to encode video data by removing redundancy while ensuring video quality as much as possible. Different from image compression tasks, video coding effectively removes redundant information in temporal domain through motion-compensated prediction techniques. Motion compensated prediction realizes the prediction of the current coding block by finding the most similar block to the current coding block in the reference frame. Therefore, improving the prediction accuracy of motion compensation is an effective way to improve the efficiency of video compression. In order to improve the co...

AI Technical Summary

Problems solved by technology

However, existing methods are mainly aimed at translational motion compensation tasks, and there is still a lack of intelligent affine motion compensation coding methods based on deep learning.

Images