论文部分内容阅读
目的高效视频编码(HEVC)采用率失真优化技术选择最佳的编码参数,实现编码比特率和视频图像失真之间的平衡。失真度量通常采用均方误差和绝对误差和,这些方法并没有考虑人眼的主观感受。为了提高视频编码的主观感知质量,提出一个融合视觉感知特性的率失真优化算法,并应用于帧间率失真优化过程中。方法首先定义了一个视觉感知因子,该因子考虑了人类视觉系统对视频图像的空域活动性区域、纹理区域、时域运动性区域和亮度的感知特性,然后以编码树单元为单位对拉格朗日乘子进行自适应调整,最后根据拉格朗日乘子与量化参数之间的关系,对量化参数做进一步的修正。结果与HEVC参考软件相比,本文算法明显提高了率失真性能,对于相同的结构相似度(SSIM)分值,本文算法在随机访问和低延时配置下平均分别节省3.1%,4.9%的码率,最高能节省9.0%的码率。与代表性文献算法相比,对于相同的SSIM,本文算法在随机访问和低延时配置下平均分别增加了0.7%,2.2%的码率节省。结论本文率失真优化策略能够根据图像不同的视觉特性自适应的调整率失真优化过程中的拉格朗日乘子,在保持编码质量基本不变的情况下,节省了码率,提高了HEVC的编码性能。
Purpose Efficient Video Coding (HEVC) uses rate-distortion optimization techniques to select the best encoding parameters to achieve a balance between encoded bitrate and video image distortion. Distortion measures usually use mean square error and absolute error sum, these methods do not consider the subjective feelings of the human eye. In order to improve the subjective perceived quality of video coding, a rate-distortion optimization algorithm based on visual perception is proposed and applied to the process of inter-frame rate-distortion optimization. The method first defines a visual perception factor, which takes into account the perceived characteristics of the human visual system in the spatial active region, the texture region, the temporal moving region and the brightness of the video image, and then uses the coding tree unit as a unit. The daily multiplier is adaptively adjusted. Finally, according to the relationship between the Lagrange multiplier and the quantization parameter, the quantization parameter is further amended. Results Compared with the HEVC reference software, our algorithm significantly improves the rate-distortion performance. For the same SSIM score, the proposed algorithm saves an average of 3.1% and 4.9% of the code respectively in random access and low latency configuration Rate, save up to 9.0% bit rate. Compared with the representative document algorithm, the proposed algorithm increases the code rate savings by 0.7% and 2.2% respectively for the same SSIM with random access and low latency configuration. Conclusion The proposed rate-distortion optimization strategy can adaptively adjust the Lagrange multipliers in the rate-distortion optimization process according to the different visual characteristics of the image, save the code rate while keeping the coding quality substantially unchanged, and improve the quality of HEVC Coding performance.