DECISION ALGORITHM FOR INTRA PREDICTION IN HIGH-EFFICIENCY VIDEO CODING (HEVC)

Prediction in HEVC exploits the redundant information fame to improve compression efficiency. The computational complexity of prediction is comparatively high as it recursively calculates depth by comparing rate-distortion optimization cost (RDO) exhaustively. deep learning technology has shown a good mark this area compared traditional signal processing because its content-based analysis and ability. This paper proposes decision algorithm predict coding tree unit (CTU) store 16-element vector, model pipelined encoder compare time taken bit rate encoding. comparison chart clearly shows reduction enhancement bitrate while dataset used here generated for with 110000 frames various resolutions, split into test, training, validation, trained on model. interfaced normal encoder. evaluation done quality check proposed BD-PSNR BD-Bitrate dip 0.6 increment 6.7 BD-Bitrate. When original HEVC, RDO an improvement over existing techniques. average encoding reduced about 72% that points complexity. An saving 88.49% achieved algorithm-based