HEVC Intra Prediction Acceleration Algorithm Based on Image Texture
Changjiang Liu, Xiaodie Zhang, Yizhong Yang · 2024
Compared to H.264/AVC, the next-generation High Efficiency Video Coding (HEVC) achieves a 50% reduction in bit rate while maintaining equivalent video quality. However, this improvement comes at the cost of significantly increased computational complexity during encoding. As a result, there is an increasing demand for fast real-time coding algorithms for HEVC encoders. To address this issue and reduce coding complexity, it is crucial to investigate efficient algorithms for HEVC. In this paper, we propose a hardware implementation algorithm that focuses on coding unit (CU) partitioning and prediction mode selection. For CU size partitioning, we introduce an adaptive algorithm based on image texture complexity to effectively filter out unnecessary coding blocks. Additionally, we optimize the selection of intra prediction modes by reducing redundant candidates based on each Prediction Unit's texture direction. This approach helps minimize computational complexity during the coding process. The experimental results demonstrate that compared with HEVC test model HM16.20, the proposed algorithm can save an average of approximately 48.89% in coding time without compromising overall performance.