3D-HEVC fast depth coding method based on Bayesian decision theorem

Abstract

The invention provides a 3D-HEVC fast depth coding method based on the Bayesian decision theorem. The method comprises the steps of firstly determining a depth map tree block, and analyzing the inter-frame mode distribution among different quantification parameters; on one hand, adopting a Bayesian rule to calculate the posterior probability of a depth map tree block, and judging whether the interaction mode of the depth map tree block is optimal to determine that the optimal coding mode of the depth map tree block is an SKIP / Merge mode; on the other hand, calculating the Bayesian cost of thedepth map tree block according to a Bayesian rule; and according to whether the Bayesian cost is less than the tolerance parameter, early determining to terminate CU trimming of the depth map tree block and determining an optimal coding mode, or segmenting the CU of the depth map tree block by using a 3D-HEVC encoder, testing other coding modes of the depth map tree block, and finding the optimalcoding mode. According to the invention, by introducing the Bayesian rule, other inter-frame modes of test depth map coding are avoided; the quality of similar virtual viewpoints is ensured; and the 3D-HEVC depth coding complexity is obviously reduced.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a 3D-HEVC fast depth coding method based on Bayesian decision theorem. Background technique [0002] In the past few years, three-dimensional (3D) video has attracted great attention through applications such as 3D movies, 3D games, and FTV. MPEG is currently developing a depth-enhanced 3D presentation called Multiview Video Plus Depth (MVD). The depth map is collocated with the texture image, the depth map represents the geometric distance between the object and the camera, and it enables the receiver to generate an intermediate view by using depth image based rendering (DIBR). To improve the compression efficiency of depth maps, JCT-3V developed 3D High Efficiency Video Coding (3D-HEVC). [0003] Depth maps are characterized differently from texture videos because depth images have large uniform regions demarcated by sharp object boundaries. Unlike texture e...

AI Technical Summary

Problems solved by technology

[0006] In view of the deficiencies in the above-mentioned background technology, the present invention proposes a 3D-HEVC fast depth coding method based on Bayesian decision theorem, which solves the technical problems of high complexity and poor rate-distortion RD performance of existing coders

Images