Analysis on Various Approaches of Complexity Reduction for Intra Prediction Modes in High Efficiency Video Coding
P. L. Chithra, Roselin Clara Angel · 2020
Video compression is an image processing technique that removes redundant data while encoding the video content. The purpose of compressing the video is to save storage space and to reduce transmission bandwidth. High Efficiency Video Coding (HEVC) is the latest compression standard available in the market and has better efficiency than its predecessor H.264/AVC. But the complexity arises during the split of the Coding Unit (CU) to form Prediction Unit (PU) and to choose the best prediction modes of the 35 prediction modes available. Hence, there is a need to reduce the computational complexity. This paper presents the different algorithms available in the literature to minimize the number of prediction modes using heuristic, machine learning and deep learning approaches. The performance analysis from the studies are tabulated and the observation made from the study is that the algorithms using deep learning techniques help to get enhanced encoding time reduction with little compromise on Bjontegaard delta bit-rate (BDBR) comparatively. The available reference software HEVC TEST Model (HM) was used as standard to study different video quality metrics such as peak signal-to-noise ratio (PSNR), Bit Rate (BD rate), and encoding time.