Depth map sequence fractal coding method based on motion vectors of color video

Abstract

The invention provides a depth map sequence fractal coding method based on motion vectors of a color video. The method includes the steps that firstly, the color video is coded through a fractal video compression method; secondly, the color video is decoded through a fractal video decompression method so as to acquire the motion vectors of all macro blocks of the color video and the motion vectors of all small blocks of the color video; thirdly, for coding of frames I in a depth map sequence, a smooth block is defined based on the H.264 intra-frame prediction coding method, and for smooth blocks, the values of adjacent reference pixels are directly copied while various predication directions do not need traversing; fourthly, for coding of frames P in the depth map sequence, block motion estimation/compensation fractal coding is carried out, predication on motion vectors of macro blocks of frames of the depth map sequence is carried out according to the correlation between the motion vectors of the macro blocks of the frames of the depth map sequence and the corresponding motion vectors of the macro blocks of the color video, an enhanced non-uniform multilevel hexagon search template is designed to replace an original non-uniform cross multilevel hexagon search template in the UMHexagonS, the most similar matching blocks are searched for through the improved UMHexagonS, and fractal parameters are recoded; fifthly, residual frames of the frames I, residual frames of the frames P and the fractal parameters of the frames P are compressed through CABAC entropy coding.

Description

Technical field: [0001] The invention belongs to the field of three-dimensional video compression coding, relates to the compression coding of depth map sequences, lays a foundation for the real-time application of three-dimensional stereo video coding, and further improves the performance of depth map-based stereo video compression coding. Background technique: [0002] The depth map represents the distance information from the scene to the camera imaging plane, and it quantizes the actual depth value to [0,255] to obtain a grayscale image representing the depth information (see Tanimoto M, Fujii T, Suzuki K, et al.Multi -view depth map of Rena and Akko&Kayo[C], ISO / IEC JTC1 / SC29 / WG11, MPEG-M14888, China: Shenzhen, 2007.). [0003] Depth map compression is the core technology in 3D video system. Currently, methods for compressing depth image sequences are mainly based on the H.264 / AVC standard. For example, Yang Xu et al. proposed a fast depth image intra-frame coding met...

AI Technical Summary

Problems solved by technology

For example, Yang Xu et al. proposed a fast depth image intra-frame coding method for the high complexity of 3D video coding. This method is based on the correlation between video images and depth images, and uses the optimal intra-frame prediction mode information of video images. Optimize the selection range of the depth image intra-frame coding prediction mode, reduce the calculation amount of the rate-distortion cost function, thereby reducing the complexity of the depth image intra-frame coding (see Yang Xu, Chang Yi, Lin Huo, etc. A fast frame of 3D video depth image Inner coding method [J]. Journal of Xidian University (Natural Science Edition), 2012,39(3):7-13.)

Images