中文摘要：

针对HEVC帧内预测planar和DC模式算法的特点,提出实现这两种模式的并行化方法。该方法是通过分析推导planar 和 DC 模式算法之间的可并行性,以PAAG（A Polymorphic Array Architecture for Graphics and Image Processing,西安邮电大学自主设计的一款面向图形、图像应用的阵列处理器）平台为基础,采用最优的数据分配方式,合理的设计了多处理单元并行工作的算法程序。实验结果表明Planar预测模式和DC预测模式在多处理单元上的并行实现,相比于单核的串行运算速度分别提高了84%和81%,串/并行加速比分别达到6.34,5.44。该并行化算法明显减少了视频的编解码时间,其数据分配方案对于帧内预测算法在多核结构上的并行化研究也有一定的参考价值。

英文摘要：

According to the algorithm characteristics of Planar and DC mode in HEVC intra prediction, a parallel processing method of the two prediction modes was proposed. The paper well designed a parallel algorithm program based on PAAG（PAAG, A Polymorphic Array Architecture for Graphics and Image Processing, an image array processor designed by Xi’an University of Posts and Telecommunications） processor platform, by analyzing and deriving the parallelism of Planar prediction mode and DC prediction mode, adopting the best data distribution, and making full use of multiple processing elements of PAAG. The experimental results show that compared with serial computing, the parallel achievement of Planar prediction mode and the DC prediction mode on the processing elements, reduced computing time of 84% and 81%, gaining speedups of 6.34 and 5.44 times, respectively. The parallel algorithm significantly reduces time video codec and the data distribution scheme has also a certain reference value for the parallelization research of intra prediction algorithm based on multi-core structure.