视频感兴趣区域快速提取与编码算法
Fast extraction and coding algorithm for video region of interest
刘鹏宇 1贾克斌1
作者信息
- 1. 北京工业大学电子信息与控制工程学院,北京100124
- 折叠
摘要
如何在复杂度和时延受限的条件下,减少时域、空域和统计冗余,获得率失真性能的最优化是视频编码设计的核心问题.作为最新的视频编码标准,H.264在取得更高压缩率的同时,编码复杂度也随之剧增,限制了其在实时环境中的应用.本文在传统视频编码框架基础上,结合人类视觉系统感知特征,提出一种基于编码信息的视频感兴趣区域快速提取算法和优化编码方案.仿真实验结果表明,本文提出的编码策略在重建视频质量基本无损失和维持原有码率的提前下,将H.264编码时间平均节省52.55％,并能够与其它快速编码技术相结合,进一步提高H.264编码速度,以更好的适应异构网络及多样性终端环境,满足日益丰富的多媒体与通信业务需求.
Abstract
How to get the Rate-Distortion Optimization under limited computation complexity and time-delay condition to reduce the time domain,space domain and statistical redundancy is the focus question of video encoding.As the newest video encoding standard,when H.264 achieves the higher coding efficiency,its computational complexity is increased acutely,and it is limited in the real-time application.Based on the conventional video encoding frame,combine the human visual system visual perception characteristics,a fast extraction algorithm for video region of interest based on the encoding information and optimized encoding scheme are proposed in this paper.The simulation results show that with almost the same quality and bit-rate compared with the H.264/AVC,the proposed method can achieve reduction of 52.55％ encoding time on average,and can be unified with other H.264 fast coding technology to enhance the encoding speed further,in order to be seasoned with heterogeneous network and multiple terminal environment,and satisfy the demand of multimedia and communication service.
关键词
H.264/视觉感知特征/编码信息/视频感兴趣区域Key words
H.264/AVC/human visual system/encoding information/video region of interest引用本文复制引用
基金项目
国家科技支撑计划项目子课题(2011BAC12B03)
国家自然科学基金面上项目(30970780)
国家青年科学基金(61100131)
北京市教委科研计划面上项目(KM201110005007)
北京市科技进步三等奖(2009)
出版年
2013
NETL
NSTL
维普
万方数据