电子学报 ›› 2018, Vol. 46 ›› Issue (3): 544-553.DOI: 10.3969/j.issn.0372-2112.2018.03.005

• 学术论文 • 上一篇    下一篇

视频压缩感知中基于结构相似的帧间组稀疏表示重构算法研究

和志杰, 杨春玲, 汤瑞东   

  1. 华南理工大学电子与信息学院, 广东广州 510640
  • 收稿日期:2016-03-15 修回日期:2016-10-07 出版日期:2018-03-25 发布日期:2018-03-25
  • 通讯作者: 杨春玲
  • 作者简介:和志杰,男,1990年生于河南漯河,华南理工大学电子与信息学院在读研究生.研究方向:视频压缩感知;汤瑞东,男,1993年生于广东汕头,华南理工大学电子与信息学院在读研究生.研究方向:视频压缩感知.
  • 基金资助:
    国家自然科学基金(No.61471173);广东省自然科学基金(No.2016A030313455)

Research on Structural Similarity Based Inter-Frame Group Sparse Representation for Compressed Video Sensing

HE Zhi-jie, YANG Chun-ling, TANG Rui-dong   

  1. School of Electronic and Information Engineering, South China University of Technology, Guangzhou, Guangdong 510640, China
  • Received:2016-03-15 Revised:2016-10-07 Online:2018-03-25 Published:2018-03-25

摘要: 基于视频帧内图像的非局部相似性和帧间信号的相关性,本文提出了一种基于结构相似的帧间组稀疏表示重构算法(SSIM-InterF-GSR),有效地提高了视频压缩感知的重构性能.在SSIM-InterF-GSR算法中,提出以结构相似度(SSIM)作为相似块匹配准则,在当前帧和参考帧内搜索匹配块生成相似块组,以相似块组的稀疏性作为正则项重构当前帧.同时,还提出了阶梯递减匹配块个数调整方案用于SSIM-InterF-GSR重构算法的迭代过程.仿真结果表明,相比于目前最好的视频压缩感知重构算法(Up-Se-AWEN-HHP),本文算法获得了更好的重构质量,最多可提升4~5dB.

关键词: 非局部相似性, 视频压缩感知, 组稀疏表示, 相似块组

Abstract: Based on the nonlocal similarity and the correlation among inter-frames in video sequences,this paper proposes an algorithm of structural similarity based inter-frame group sparse representation(SSIM-InterF-GSR),which effectively improves the reconstruction performance for compressed video sensing.In SSIM-InterF-GSR,the structural similarity(SSIM)is utilized as block matching criterion to generate the group of similar blocks from the current frame and reference frames.And then,the sparsity of the groups is used as the regularization term to reconstruct the current frame.Meanwhile,the step-decreasing scheme for number of matching blocks is proposed during the iteration process of SSIM-InterF-GSR.Simulation results show that,compared to the state-of-the-art compressed video sensing reconstruction algorithm(Up-Se-AWEN-HHP),the SSIM-InterF-GSR algorithm obtains a better reconstruction quality.The most gap is up to 4~5dB.

Key words: nonlocal similarity, compressed video sensing, group-based sparse representation, the group of similar blocks

中图分类号: