Method for hiding loss errors of three-dimensional video macro blocks

Abstract

The invention discloses a method for hiding the loss errors of three-dimensional video macro blocks. According to the method, the motion and static states of the macro blocks in a lost frame are estimated by prejudging the magnitude of time domain relevance between adjacent time domain frames of the lost frame and adjacent frames among viewpoints, the macro blocks are classified into motion macro blocks and static macro blocks, and lost macro blocks are restored by adopting a direct time domain frame copying method specific to static macro blocks with high time domain relevance, so that the calculation complexity is low, and the subjective and objective quality of the type of lost macro blocks can be improved; and specific to the motion macro blocks, an optimal candidate macro block is selected from candidate macro blocks pointed by each vector in a candidate vector set of the lost macro blocks by adopting two boundary matching principles, optimally matched macro block recovery or multi-direction interpolation method recovery is adaptively selected according to the boundary distortion value of the optimal candidate macro block, so that the subjective and objective quality of the lost macro blocks moving vigorously can be improved, and a very good recovering effect can be achieved on the lost macro blocks positioned on the edge of an object.

Description

technical field [0001] The invention relates to a video error concealment method, in particular to a stereoscopic video macro block loss error concealment method. Background technique [0002] With the rapid development of three-dimensional display and related technologies, three-dimensional (stereoscopic) video is more and more valued by people because it can provide a new three-dimensional experience and meet the viewers' needs for depth, interaction and three-dimensionality. . With the wide application of video communication technology, how to effectively realize high-quality and stable video transmission has become a higher goal pursued by researchers. In the process of stereoscopic video coding and transmission, due to the large amount of data, stereoscopic video coding methods use spatial domain, time domain and viewpoint domain prediction to eliminate redundant information, which reduces the anti-error performance of the transmission code stream and affects the chann...

AI Technical Summary

Problems solved by technology

For stereoscopic video, the strength of temporal correlation in different regions is different. For the lost macroblocks with strong temporal correlation, using the temporal frame copy method can not only achieve better restoration effect, but also has low computational complexity. , so it is unreasonable to adopt the same concealment method for the lost macroblocks in different regions

The existing stereoscopic video macroblock loss error concealment methods do not take into account the characteristics of missing macroblocks in different regions and the temporal and spatial domain correlation and inter-view correlation of error frames.

Images