Abstract: 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.

Key words: H.264/AVC standard, Human Visual System(HVS), motion vector, inter-frame prediction mode, Region of Interest(ROI)

摘要： 针对H.264/AVC视频编码标准计算复杂度高、难以实时应用的缺陷，利用重建块的运动矢量和帧间预测模式，并结合人类视觉感知特征，提出一种视频感兴趣区域快速提取算法及其优化编码方案。仿真结果表明，与H.264/AVC算法相比，该算法能够在码率平均增加0.17%、峰值信噪比平均下降0.052 dB的前提下，节省53.11%的编码时间。

关键词: H.264/AVC标准, 人类视觉系统, 运动矢量, 帧间预测模式, 感兴趣区域

CLC Number: 

• TP391