论文部分内容阅读
在H.264/AVC视频编解码标准中,帧内预测利用周围像素预测当前块来降低空间冗余,能极大地提高H.264/AVC的编码效率。然而,帧内预测的4×4预测有9种模式,为选出最佳模式,全搜索算法需要花费极大的计算量。为了降低帧内预测的复杂度,文章结合9种预测模式的各自特点和整数变换的线性性质,提出一种适用于H.264/AVC帧内4×4模式选择的快速变换和量化算法来减少对应模式变换和量化操作步骤的计算量。实验结果表明,优化算法可以有效地减少整数变换和量化操作的计算量。
In the H.264 / AVC video codec standard, intra prediction uses the surrounding pixels to predict the current block to reduce spatial redundancy, which can greatly improve the coding efficiency of H.264 / AVC. However, there are 9 modes for 4 × 4 prediction of intra prediction, and full search algorithm takes a great amount of computation to select the best mode. In order to reduce the complexity of intra prediction, this paper proposes a fast transform and quantization algorithm suitable for 4 × 4 mode selection in H.264 / AVC frame to reduce the complexity of the 9 prediction modes and the linearity of integer transform Corresponds to the pattern transformation and quantification of calculation steps. Experimental results show that the optimization algorithm can effectively reduce the computational complexity of integer transform and quantization operations.