Hardware-Oriented Hierarchical Motion Estimation Algorithm for HEVC

Motion estimation is one of the most computational complex parts of HEVC. In order to improve the coding efficiency of HEVC compressing UHD video, this paper adopts the method of combining central full search and peripheral downsampling search into the first layer, and a hexagon fast search in the second layer to optimize the first layer search results. After describing the proposed algorithm in the latest HEVC software model (HM 15.0), the simulation test for the UHD video sequence shows that the proposed method can reduce the computational complexity of the full search motion estimation algorithm by 97.8%, while the BD-rate loss is only 0.77%, which shows that the proposed algorithm can achieve efficient encoding.