欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2018, Vol. 35 ›› Issue (3): 425-431.DOI: 10.7523/j.issn.2095-6134.2018.03.018

• 简报 • 上一篇    

简报基于感兴趣区域的HEVC会话视频容错编码新方案

和智涛, 张灿, 陈德元   

  1. 中国科学院大学电子电气与通信工程学院, 北京 101408
  • 收稿日期:2017-05-04 修回日期:2017-05-31 发布日期:2018-05-15
  • 通讯作者: 张灿
  • 基金资助:
    国家自然科学基金(61571416,41776204)和中国科学院大学校所合作基金(Y65201GY00)资助

Region-of-interest based error resilient scheme for HEVC conversational video transmission

HE Zhitao, ZHANG Can, CHEN Deyuan   

  1. School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 101408, China
  • Received:2017-05-04 Revised:2017-05-31 Published:2018-05-15

摘要: 在计算和传输资源受限的智能手机上编码传输实时会话视频一直是引人关注的关键技术。在采用肤色、运动和纹理信息快速提取感兴趣区域并且使用丢包率修正感兴趣区域权重的基础上,提出一种HEVC会话视频容错编码方案。仿真结果表明,该方案在丢包率5%~20%的网络中编码传输时,会话视频的感兴趣区峰值信噪比PSNR比现存的随机帧内刷新方法至少提高0.7 dB,编码时间减少20%~40%。实际比特率符合目标比特率要求。

关键词: 会话视频传输, 感兴趣区, 高效率视频编码, 人类视觉系统

Abstract: Conversational video coding and transmission on resource-constrained mobile phones has been a key technology. We propose a novel error resilient scheme for conversational video transmission in HEVC which is based on the fast extraction for region-of-interest (ROI) according to skin color, motion, and texture information and adjustments for the ROI weights using the packet-loss rate. The simulation results show that, compared to the random frame refresh method, the proposed scheme improves PSNR of the ROI by at least 0.7dB and reduces coding time by 20%-40% when the video coding and delivery is over the channel with the packet-loss rate of 5%-20%. At the same time the actual bitrate matches the target bitrate.

Key words: conversational video transmission, region of interest, HEVC, human visual system

中图分类号: