论文部分内容阅读
为降低空时分级视频编码中自适应层间预测模式的计算复杂度,利用增强层中时域宏块相关性的特点,提出一种随时间等级变化的快速模式决策算法。该算法将参考宏块使用过的模式按照其率失真优化代价排序,选择其中代价较小的模式作为当前宏块的动态备选方案,从而有效地提高编码效率。在JSVM上的实验结果表明,同原有非优化决策算法相比,在PSNR平均降低0.0148dB、比特率增加0.55%的情况下,该优化算法使得平均编码时间减少21.67%。
In order to reduce the computational complexity of adaptive inter-layer prediction mode in space-time hierarchical video coding, a fast mode decision algorithm with temporal level changes is proposed by using the characteristics of time-domain macroblock correlation in enhancement layer. The algorithm uses the reference macroblock as a model to optimize the cost according to the rate-distortion optimization, and selects the less-expensive mode as the dynamic alternative for the current macroblock to effectively improve the coding efficiency. Experimental results on JSVM show that the proposed algorithm can reduce the average coding time by 21.67% when the average PSNR decreases 0.0148dB and the bit rate increases 0.55% compared with the original non-optimal decision algorithm.