Video ROI Extraction Algorithm Based on Reconstructed Encoding Information

Aiming at defects of the high computational complexity and hardly used in real-time application of H.264/AVC video encoding standard,by using reconstruction of macro-blocks motion vector and inter-frame prediction mode,and combining feature of Human Visual System(HVS),this paper proposes a fast extraction algorithm for video Region of Interest(ROI) and optimized encoding scheme.Simulation results show that compared with H.264/AVC algorithm,the proposed algorithm can achieve reduction of 53.11% encoding time on average,while PSNR only declined 0.052 dB and bit-rate increase 0.17% at the same time.