摘要
详细分析了统一计算设备架构(CUDA)的编程模型,从并行计算角度对H.264视频编解码中的去方块滤波进行研究和优化,提出了基于CUDA加速的去方块滤波并行处理方法。通过对高清测试序列的实验表明,利用GPU并行处理能力能够明显提高视频编解码速度,并有效降低CPU资源占用率。
The programming model of Compute Unified Device Architecture (CUDA) is analyzed, and the algorithm of deblocking filter in H.264 from the aspect of the parallel computing is studied. Based on CUDA, design and implementation of deblocking filter with the method of parallel computing are presented. Experimental results of high-definition video show that, with the assistance of GPU, the processing speed of video encoding and decoding is improved, and the CPU resource utilization is reduced effectively.
出处
《电视技术》
北大核心
2010年第5期44-46,共3页
Video Engineering