基于熵平衡的全景视频编码  

Panoramic video coding based on entropy balancing

作　　者：杨栩 郭红伟[3] 李婉雪 Yang Xu;Guo Hongwei;Li Wanxue(School of Physics&Engineering Technology,Chengdu Normal University,Chengdu 611130,China;School of Information&Communication Engineering,Chongqing University of Posts&Telecommunications,Chongqing 400065,China;College of Engineering,Honghe University,Honghe Yunnan 661199,China)

机构地区：[1]成都师范学院物理与工程技术学院,成都611130 [2]重庆邮电大学通信与信息工程学院,重庆400065 [3]红河学院工学院,云南红河661199

出　　处：《计算机应用研究》2023年第6期1894-1899,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(62061015);成都师范学院专项科研项目(ZZBS2020-11);云南省基础研究计划资助项目(202101AT070051)。

摘　　要：高效的全景视频压缩对沉浸式、低延时的视觉体验十分重要。服务器端的全景视频通过投影到2D平面,经过编码、传输到客户端后,反投影到球面供用户观看,而全景视频投影过程中会在两极产生拉伸形变,由于过采样而产生像素冗余,破坏原有的信息熵分布。针对这一问题,提出一种改进的熵平衡算法,根据全景视频投影前后对应编码单元的熵不变原则,采用码率控制的R-λ模型根据投影面积变化比率来重新分配目标比特,以达到熵平衡,进而调整编码参数优化编码。实验结果表明,该算法相较于基准VTM14.0能达到1.87%的码率节省,编码时间下降10.07%,说明该算法在全景视频编码中能实现更高的压缩率和更短的编码时间。Efficient panoramic video compression is very important for immersive and low-latency visual experience.The panoramic video on the server is projected onto the 2D plane,encoded and transmitted to the client and then back-projected the 2D image to the spherical surface for users to watch.However,The panoramic video projection will produce stretching deformation in the two poles and pixel redundancy,and destroy the original information entropy distribution.Aiming at this problem,this paper proposed an improved entropy balance algorithm.To achieve entropy balance,this paper used the R-λmodel in rate control to redistribute the target bit according to the projection area change ratio,and then adjusted the coding parameters to optimize the coding.The experimental results show that the proposed algorithm can achieve 1.87%bit rate saving and 10.07%coding time reduction compared with the benchmark VTM14.0,and indicate that the algorithm can achieve higher compression rate and shorter coding time in panoramic video coding.

关 键 词：全景视频 熵平衡 ERP投影 率失真优化 VTM 

分 类 号：TN919.81[电子电信—通信与信息系统]