Abstract
The High Efficiency Video Coding (HEVC) is the latest video compression standard developed by JCT-VC. It adopts a flexible quad-tree structured coding unit (CU). Each CU can be recursively split into four sub-CUs of equal size. At each CU depth level, various combinations of block partitioning types and prediction modes are evaluated to find the one with the least rate distortion (RD) cost. This introduces very high computation complexity. To reduce the encoding time, we propose three early termination schemes for eliminating unpromising splitting. In the first algorithm, calculated RD costs of the parent CU and sibling CUs of the current CU are used to avoid unnecessary CU splitting. In the second scheme, an additional threshold value is employed to terminate the splitting process in advance. In the third approach, we utilize the correlationship between RD costs and prediction modes and an adaptive threshold value to make an even earlier termination decision. Experimental result shows that the proposed algorithm reduces averagely 15.7% of encoding time with negligible quality loss nor bit-rate overhead.