Method for concealing entire frame loss error of multi-view video

Abstract

The invention discloses a method for concealing an entire frame loss error of multi-view video. The method comprises the following steps: prejudging the strength of the time domain relevance and the inter-view relevance of all image blocks in a lost frame so as to select a proper reference frame; recovering the image blocks in the lost frame by adopting different recovery methods; and performing hole area filling on an initial recovered frame obtained by recovery to obtain a final recovered frame of the lost frame. The method for concealing the entire frame loss error of the multi-view video has the advantages that the time domain relevance and the inter-view relevance are effectively utilized, and for the image blocks prejudged to have stronger time domain relevance in the lost frame, the image blocks are recovered by adopting a method of direct time domain copy, and thereby, not only can the computational complexity be reduced, but also the subjective and object quality of each recovered image block can be increased, whereas for the image blocks prejudged to have lower time domain relevance in the lost frame, the image blocks are recovered by adopting a method of inter-view predictive recovery, and thereby, the accuracy of recovering the image blocks can be effectively increased.

Description

technical field [0001] The invention relates to a video error concealment method, in particular to a multi-viewpoint video full frame loss error concealment method. Background technique [0002] Multi-viewpoint video can be widely used in free-viewpoint TV, 3D TV, and the like. Compared with single-view video, multi-view video transmits a huge amount of data, so it must be effectively compressed and encoded. Almost all current multi-view video coding schemes use temporal and spatial prediction to eliminate redundant information in time domain, spatial domain and between views, making the encoded multi-view video stream very sensitive to channel errors. However, due to channel congestion and other reasons, network packet loss occurs from time to time. Once an error occurs, if it is not recovered, it will not only affect the correct decoding of the current erroneous frame and subsequent frames of the current view, but also because the multi-view video coding uses disparity es...

AI Technical Summary

Problems solved by technology

However, for the case of whole frame loss, because there is no correctly decoded macroblock in the damaged frame, the existing error concealment algorithm for partial macroblock loss cannot be directly applied to the case of multi-view video full frame loss; On the other hand, for macroblocks located in areas with severe motion, there is often a large difference between the motion of the damaged macroblock and the macroblock at the same position in the reference frame, which leads to inaccurate prediction and affects the restoration quality of the damaged block; in addition , the existing error concealment algorithm for the loss of some macroblocks does not fully determine the optimal recovery method when calculating several candidate blocks of the damaged macroblock. When restoring the lost macroblock, it needs to restore multiple times to obtain different candidate blocks. lead to increased computational complexity

Images