基于视觉注意的视频可伸缩ROI算法

山东大学学报(工学版) ›› 2013, Vol. 43 ›› Issue (1): 15-21.

基于视觉注意的视频可伸缩ROI算法

陈旭1, 张基宏2, 柳伟2, 梁永生2

1. 深圳大学信息工程学院, 广东深圳 518060;
2. 深圳信息职业技术学院可视媒体处理与传输深圳市重点实验室, 广东深圳 518029

收稿日期:2012-12-05 出版日期:2013-02-20 发布日期:2012-12-05
作者简介:陈旭(1984- )，男，广东广州人，博士研究生，主要研究方向为自适应可伸缩视频编码. E-mail:chenxu-email@163.com
基金资助:
国家自然科学基金资助项目(60772163); 广东省自然科学基金资助项目(S2011010006113); 深圳市国际科技合作资助项目(ZYA201007070116A)

New scalable ROI algorithm based on visual attention

CHEN Xu1, ZHANG Ji-hong2, LIU Wei2, LIANG Yong-sheng2

1. School of Information Engineering, Shenzhen University, Shenzhen 518060, China;
2. Shenzhen Key Lab of Visual Media Processing and Transmission, Shenzhen Institute of Information Technology,Shenzhen 518029, China

Received:2012-12-05 Online:2013-02-20 Published:2012-12-05

摘要/Abstract

摘要：

如何合理进行码率分配达到充分利用码率资源目的,提高视频的主观视觉质量一直是视频编码领域研究的关键问题。本研究首先提出一种基于人眼视觉注意检测模型,该模型的核心思想是针对人眼不同时间段内进行眼动策略分析;然后结合显著计算得到图像注视区域结果得到SVC增强层码率分配算法。实验测试结果表明本研究提出的算法与现有其他算法相比具有更为有效的模拟人眼视觉注意的特性,在有限网络带宽条件下使得人眼视觉效果得到显著提升。

关键词: 显著计算, 码流提取, H.264／SVC, 视觉注意, 眼动

Abstract:

In order to make full use of video coding bit-steam and high subjective QoE(quality of experience), a AOF(area of fixation) detection model was proposed based on current achievements of psychology and visual attention. Combined with saliency calculation and eye-movement tracking analysis, this model could obtain the AOF of different viewing duration. According to the AOF, a method of content-aware rate adaptation was also proposed. The experimental results showed that the proposed method could outperform than the other existing methods and could improve the subjective QoE when coding and transmission in situation of limited bandwidth.

Key words: H.264／SVC, eye movement, visual attention, saliency calculation, bitstream extractor

中图分类号:

TP391

陈旭1, 张基宏2, 柳伟2, 梁永生2. 基于视觉注意的视频可伸缩ROI算法[J]. 山东大学学报(工学版), 2013, 43(1): 15-21.

CHEN Xu1, ZHANG Ji-hong2, LIU Wei2, LIANG Yong-sheng2. New scalable ROI algorithm based on visual attention[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2013, 43(1): 15-21.

参考文献

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed