一种HEVC帧内快速编码算法

发布时间：2018-04-12 07:03

本文选题：高效视频编码(HEVC) + 编码单元(CU)　；参考：《光电子·激光》2017年03期

【摘要】：高效视频编码(HEVC)采用编码单元(CU)四叉树的分割结构,相比H.264/AVC显著地提升了编码效率,但却使编码复杂度急剧增加。为此,本文提出一种帧内快速编码算法。首先,根据视频图像纹理复杂度,提前判断是否进行最大编码单元(LCU)分割。然后,根据空域相邻CU的深度预测当前CU的深度范围,跳过不必要的计算;最后,根据预测模式被选为最优预测模式的统计特性,去掉可能性小的帧内预测模式。本文算法在HM14.0的基础上实现。仿真结果表明,本文算法在全I帧模式下与HM14.0相比,帧内编码时间平均减少38%,码率(BR)只增加1.41%,峰值信噪比(PSNR)只降低0.29dB,在保证编码性能和视频质量几乎不变的情况下,本文算法降低了编码的计算复杂度。
[Abstract]:The efficient video coding (HEVC) is based on the partition structure of the coding unit CU4 (Quadtree). Compared with H.264/AVC, the coding efficiency is significantly improved, but the coding complexity is increased dramatically.Therefore, this paper proposes a fast intra-frame coding algorithm.Firstly, according to the texture complexity of video image, the maximum coding unit (LCU) segmentation is determined in advance.Then, the depth range of the current CU is predicted according to the depth of adjacent CU in the airspace, and the unnecessary calculation is skipped. Finally, according to the statistical characteristics of the optimal prediction mode, the intra prediction mode, which is less likely to be possible, is removed.This algorithm is implemented on the basis of HM14.0.The simulation results show that compared with HM14.0 in full I frame mode, the coding time in this algorithm is reduced by 38, the bit rate is increased by only 1.41%, and the PSNR (peak signal-to-noise ratio) is reduced by 0.29 dB, while the coding performance and video quality are almost unchanged.The algorithm reduces the computational complexity of the coding.
【作者单位】：重庆邮电大学光通信与网络重点实验室;
【基金】：国家自然科学基金(61371096,61171158,61571072) 重庆市基础与前沿研究计划(cstc2015jcyjA40015)资助项目
【分类号】：TN919.81

【相似文献】