A 3D video intelligent multi-domain joint predictive coding method and device

Abstract

The invention discloses a 3D video intelligent multi-domain joint predictive coding method and device, including: 1) Obtaining multi-domain reference information: reconstructing the left side, the upper side and the upper left side of the current coding block within the range of step size As spatial reference information; use the inter-frame prediction block of temporal correlation of adjacent frames as temporal reference information; use the view synthesis prediction block obtained through view synthesis prediction technology as inter-view reference information; 2) build a spatio-temporal prediction network to The spatio-temporal domain reference information is used as input to obtain the spatio-temporal domain prediction result; 3) A multi-domain joint prediction network is constructed according to the spatio-temporal domain prediction result and the viewpoint synthesis prediction block to obtain the final multi-domain prediction result. The device includes: a memory, a processor, and a computer program stored in the memory and operable on the processor. The processor implements the steps of the method when executing the program.

Description

technical field [0001] The present invention relates to the field of 3D video coding, in particular to a 3D video intelligent multi-domain joint predictive coding method and device. Background technique [0002] With the development of 3D technology, 3D video coding has become a major research hotspot in the field of multimedia. Compared with 2D video, 3D video has more data volume, which brings great challenges to video storage and transmission. Therefore, how to realize efficient 3D video compression coding has important theoretical research significance and practical application value. [0003] As a new generation video coding standard, HEVC (High Efficiency Video Coding) effectively improves compression efficiency. As a 3D extension of HEVC, 3D-HEVC adopts a coding architecture based on the MVD (Multiview Videoplus Depth, multi-viewpoint plus depth) video format. Based on HEVC’s existing technology, it adds a new technology for multi-viewpoint video and depth video cod...

AI Technical Summary

Problems solved by technology

[0006] The invention provides a 3D video intelligent multi-domain joint predictive coding method and device. The invention comprehensively analyzes and mines the time domain, space domain and viewpoint domain correlation of 3D video, proposes to use CNN to fuse multi-domain reference information, and proposes a hierarchical A multi-domain prediction mechanism is used to solve the problem of multi-domain reference information fusion; in addition, in the hierarchical prediction mechanism, an effective multi-domain joint prediction network is constructed, and a multi-scale coding unit is designed in the network to extract features. Use CNN to solve the multi-domain joint prediction problem of 3D video, see the description below for details:

Images