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球面的两极产生拉伸形变,进而产生像素冗余.针对这一问题,提出一种自适应量化参数算法,根据全景视频纬度方位来调整编码参数优化编码.实验结果表明,提出的算法相较于基准VTM14.0能达到1.7%的码率节省,编码时间下降5.56%.
Efficient panoramic video compression is important for an immersive,low latency visual experience.The panoramic video on the server side is projected onto a 2D plane,encoded,transmitted to the client and then back-projected onto a sphere for the user to view.The panoramic video projection process,however,generates stretching deformation at the poles,which in turn generates pixel redundancy.To address this problem,an adaptive quantization parameter algorithm is proposed to optimise the coding parameters according to the latitude orientation of the panoramic video.Experimental results show that the proposed algorithm achieves a 1.7%code rate saving and a 5.56%reduction in coding time compared to the benchmark VTM14.0.
作者
杨栩
YANG Xu(School of Physics and Engineering Technology,Chengdu Normal University,Chengdu,Sichuan 611130)
出处
《绵阳师范学院学报》
2023年第8期6-12,共7页
Journal of Mianyang Teachers＇ College
基金
成都师范学院专项科研项目(ZZBS2020-11).
关键词
全景视频
量化参数
ERP投影
率失真优化
VTM
Panoramic video
Quantification parameter
ERP projection
Rate distortion optimization
VTM
