Abstract:

3D extension of high efficiency video coding (3D-HEVC) has become a video compression international standard for multi-view video system applications. The mode prediction process must search for optimal prediction modes and depth levels with a minimum rate-distortion cost for the coding unit (CU). New coding techniques with improved texture video coding efficiency have been introduced; however, they also increase computational burden, which obstructs 3D-HEVC encoder reliability applications. Therefore, the development of a novel texture video coding is necessary to reduce 3D-HEVC computational complexity. In this paper, the proposed algorithm will exploit an adjusted method of DE and ME based on motion homogeneity of texture videos. The algorithm includes early SKIP/Merge mode decision and adaptive motion search range adjustment. As shown by the simulation results, the proposed fast texture video coding method can reduce computational complexity by approximately 56.6% on average while resulting in a tiny loss of quality.