A Fast Full Search Scheme for Rate-Distortion Optimization of Variable Block Size and Multi-frame Motion Estimation

Two new fast full search (FFS) methods are proposed which can improve the performance of the rate- distortion optimized video encoding process. The first method combines the concepts of both conventional sum of absolute difference (SAD) reuse method and row-based partial distortion search (PDS) together to make the performance largely better than these two conventional methods for variable block size motion estimation. The second method speeds up the multi-frame motion estimation process by using the previous results obtained in considered reference frames to set extra thresholds for rejecting search points sooner in reference frames. Furthermore, these two proposed methods can be easily combined to form a very effective motion compensation scheme. On average, a 62% and 52% reduction in terms of computational complexity of integer variable block size and multi-frame motion estimation can be achieved without any video quality degradation as compared with that of the conventional PDS and conventional SAD reuse method respectively.