High Efficiency Video Coding (HEVC) employs the quadtree coding tree unit (CTU) structure to improve the coding efficiency. However, to achieve the coding efficiency, all possible block partition should be performed on the each depth level to obtain best combination of the coding unit (CU), prediction unit (PU), and transform unit (TU). To reduce the complexity of block partitioning process, this thesis proposed a fast CU decision algorithm that includes PU mode skip scheme with motion vector based Region of Interest CTU selection and depth level skip scheme based on the co-located block information in the previously coded frames. The RoI selection distinguishes between a dynamic region CTU and a static region CTU to decide whether perform the finer block partitioning or not. The depth level skip scheme performs depth averaging process and estimate the most probable target depth level. The experimental results show that the proposed fast CU decision algorithm can reduce the total encoding time up to 55%, with compared to the original HEVC HM reference software encoder. Moreover, the PSNR and BDBR loss of the proposed algorithm is negligible while gaining a substantial reduction in terms of the encoding time.
