A fast mode decision algorithm for multiview video coding

Abstract

In this paper, a fast mode decision algorithm for multi-view video coding (MVC) is proposed. When implementing MVC, one of the most critical problems is heavy computational complexity caused by the large amount of information in multi-view sequences. Hence, a fast algorithm is essential. To reduce computational complexity, a fast mode decision using both motion vector among views and texture characteristic is proposed to reduce encoding time. For MBs with homogeneous motion and texture, only mode of Skip and Inter16×16 are tested, and other mode sizes are skipped; for MBs with active motion or rich texture, all mode sizes are tested. Experimental results showed that compared with the full mode decision (FMD) in JMVM, the proposed algorithm is able to reduce processing time by maximumly 60–70% in estimation process, and maintain almost the same coding efficiency.