IEIE Transactions on Smart Processing and Computing, vol. 5, no. 4, August 2016 http://dx.doi.org/10.5573/IEIESPC.2016.5.4.231 231 IEIE Transactions on Smart Processing and Computing Depth-of-interest–based Bypass Coding–unit Algorithm for Inter-prediction in High-efficiency Video Coding Chae Eun Rhee Department of Information and Communication, Inha University / Incheon, South Korea chae.rhee@inha.ac.kr Received June 27, 2016; Accepted July 21, 2016; Published August 30, 2016 * Short Paper Abstract: The next-generation video coding standard known as High-Efficiency Video Coding (HEVC) was developed with the aim of doubling the bitrate reduction offered by H.264/Advanced Video Coding (AVC) at the expense of an increase in computational complexity. Mode decision with motion estimation is still one of the most time-consuming computations in HEVC, as it is with H.264/AVC. Several schemes for a fast mode decision have been presented in reference software and in other studies. However, a possible speed-up in conventional schemes is sometimes insignificant for videos that have inhomogeneous spatial and temporal characteristics. This paper proposes a bypass algorithm to skip large-block-size predictions for videos where small block sizes are preferred over large ones. The proposed algorithm does not overlap with those in previous works, and thus, is easily used with other fast algorithms. Consequently, an independent speed-up is possible. Keywords: HEVC, Inter-prediction, Fast mode decision 1. Introduction Recently, the next-generation video coding standard [1- 3] known as High-Efficiency Video Coding (HEVC) was developed by the ISO/IEC Moving Picture Experts Group (MPEG) and the ITU-T Video Coding Experts Group (VCEG). In the emerging HEVC standard, several new features were introduced. In particular, various block sizes from 8×8 to 64×64 and flexible block size selection schemes for predictions and transforms have been shown to be very suitable for large resolutions. Thus, mode decisions with motion estimation (ME) are among the most time-consuming computations in HEVC. HEVC reference software and other recent studies suggest several schemes for rapid mode decisions. These algorithms are useful for speed-up in most cases. However, there can be very little benefit for videos that have inhomogeneous spatial and temporal characteristics. To make up for this weakness, there are numerous ongoing studies for HEVC fast mode decisions that pursue an effective trade-off between degradation of compression efficiency and speed-up degree [4-7]. This paper proposes fast algorithms that attempt to determine the block size early, and consequently, speed up the encoding time by avoiding computations with unselected block sizes. Depending on the video content, large and small blocks have a different probability of being selected as the optimal block size. To skip large blocks for videos that have complex motion and texture characteristics, an intra-prediction routine is executed first. If the rate-distortion (R-D) cost of intra-prediction in the current block size is small enough, compared to that of inter-prediction, smaller block sizes are expected to be more suitable than the current selection due to the high spatial complexity. Thus, the current block size is bypassed and searches continue on smaller block sizes. The fast decision points of the proposed algorithms do not overlap those proposed in the HEVC reference software or in previous studies. Thus, the proposed algorithms are easily used with other fast-decision algorithms, and the resulting speed-up effect is quite independent. Simulation results show that the encoding speed is improved by over 10% for videos that have complex texture and motion, whereas the R-D degradation is marginal. The rest of this paper is organized as follows. Section 2 gives previous fast-decision algorithms. The proposed algorithm and simulation results are presented in sections 3 and 4, respectively. Conclusions are given in Section 5.