METHOD FOR ACQUIRING ADJACENT DISPARITY VECTORS IN MULTI-TEXTURE MULTI-DEPTH VIDEO

摘要

The present invention relates to the technical field of 3D-HEVC video coding. Disclosed is a method for acquiring adjacent disparity vectors in auxiliary viewpoint coding in multi-texture multi-depth video coding. The method is characterized in that a criteria of taking a firstly found disparity vector as a final disparity vector is changed, and comprises: grouping adjacent candidate spatial and temporal coding unit positions by deleting least searched positions from candidate spatial and temporal coding unit positions adjacent to a current coding unit; and combining the found disparity vectors thereof according to an adoption ratio into a final disparity vector. The method improves a coding quality while retaining an efficiency of an original fast algorithm, addresses a lack of coding quality of an original method in which the number of search times is reduced by deleting search candidate positions, and improves the coding quality by at least 0.05% while retaining the efficiency of a present fast algorithm (decoding time is reduced to 97.1% of original decoding time, and coding and virtual viewpoint synthesis time is unchanged).