论文部分内容阅读
提出了一种基于宏块级运动预检测的模式选择算法.采用低复杂度的联合运动检测准则对当前宏块运动程度进行评估.根据当前宏块与空间相邻宏块的运动程度将宏块分级,并采用不同的编码模式判决方法.该算法能较好地区分背景噪声与运动物体,并尽可能保留运动宏块细节.对一些典型监控场景视频序列的仿真实验结果显示,该算法平均节约了75.2%的编码时间.与H.264参考软件中的模式选择算法相比,该算法不但节约了1.31%的平均码率,而且平均峰值信噪比提高了约0.08dB.
This paper proposes a mode selection algorithm based on macroblock-level motion pre-detection.Using the joint motion detection criterion with low complexity, the motion of the current macroblock is evaluated.According to the motion of the current macroblock and the adjacent macroblock, the macroblock The algorithm can distinguish the background noise and the moving object well, and retain the details of the moving macroblock as possible.A simulation experiment on some typical monitoring scene video sequences shows that the algorithm saves the average Up to 75.2% coding time.Compared with the mode selection algorithm in H.264 reference software, this algorithm not only saves the average bit rate of 1.31%, but also improves the average peak signal-to-noise ratio by about 0.08dB.