Transform-based fast mode and depth decision algorithm for HEVC intra prediction

The High Efficiency Video Coding (HEVC), as the new standard of video coding, provides significantly better coding efficiency than all existing video coding standards. One key contributor to this improvement is the new intra prediction method that supports a large number of prediction directions in the quad-tree structure. It takes a very high cost of computational complexity. This paper presents a transform-based fast mode and depth decision algorithm for HEVC intra prediction. Based on the inherent features of transform, the block is analyzed in the frequency domain to determine its directional tendency. About two thirds of modes are removed. Moreover, using the information of detected directions and the transformed coefficients, we reduce the calculated depths from 5 to 2.4. Experimental results show that the proposed method provides a decrease of 60.3% computation time, compared to HEVC reference software, with only 2.18% in the bit-rate increment.