论文部分内容阅读
为减轻质量域可分级视频编码的计算复杂度,提出了一种快速模式选择算法.该算法将模式选择流程分为3个阶段:首先检测运算复杂度低、出现概率大的跳过类模式集合;依据层间相关性和空域相关性选择性检测候选编码模式集合;最后遍历检测剩余编码模式.若第1或第2阶段所获得的最小码率-失真代价值小于预先定义的门限,则提前结束模式选择流程.仿真结果证明,与全遍历算法相比,新算法可以获得45.9%～80.8%的编码速度提升;峰值信噪比降低最多为0.15 dB,码率增加最多为0.62%.
In order to reduce the computational complexity of scalable video coding in the quality domain, a fast mode selection algorithm is proposed, which divides the mode selection process into three stages: Firstly, it detects the set of skip modes with low computational complexity and high probability of occurrence ; Selectively detect the set of candidate encoding modes according to the correlation between layers and the spatial correlation; and finally traverse the remaining encoding modes, and if the minimum rate-distortion cost value obtained in the first or the second stage is less than the predefined threshold, advance The simulation results show that the new algorithm can achieve the coding speed improvement of 45.9% -80.8% compared with the full traversal algorithm, the peak signal-noise ratio is reduced by up to 0.15 dB, and the rate increase is up to 0.62%.