GA-Based motion optimization for overlapped block motion compensation

Overlapped block motion compensation (OBMC) provides improved prediction efficiency as well as better subjective quality for coded video compared with conventional block motion compensation (BMC). However, for high-activity sequences, the estimation results for OBMC are not significantly improved because the motion vectors are not sufficiently optimized. In this paper, We present a new approach to motion optimization for OBMC. The motion vectors are optimized using a Genetic Algorithm (GA) that is robust and effective to deal with complex combinatorial problems. The improvements with respect to existing methods have been confirmed by experimental results.