基于PCA硬阈值收缩的平滑投影Landweber图像压缩感知重构

被引:10
作者
李然
干宗良
朱秀昌
机构
[1] 南京邮电大学江苏省图像处理和图像通信重点实验室
关键词
分块压缩感知; 边缘检测; 自适应测量; 主成分分析; 硬阈值收缩; 方向小波; 平滑投影Landweber重构;
D O I
暂无
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
利用压缩感知理论对图像进行测量和重构时,基于分块思想可有效提高重构速度,但同时会带来较强的块效应。为了解决该问题,在编码端提出了一种基于边缘检测的自适应分块压缩感知测量方案;在解码端提出了一种基于主成分分析(PCA)的平滑投影Landweber(SPL)重构法,该算法运用PCA训练出适合于图像结构的稀疏字典,用于进行硬阈值收缩,从而有效消除了块效应,提升了重构图像的质量。为了提高硬阈值收缩效率和减少训练复杂度,采用了3种基于块的PCA硬阈值收缩方案:全局PCA、局部PCA和分层PCA。仿真实验结果表明:所提出的自适应压缩感知测量方案与SPL重构法相结合,和传统分块压缩感知方案相比,峰值信噪比(PSNR)值均提升了1~3 dB;本文算法,无论在传统分块压缩感知方案下还是在自适应分块压缩感知方案下,与基于方向小波阈值收缩的SPL重构算法相比,均获得了更高的PSNR值。
引用
收藏
页码:504 / 514
页数:11
相关论文
共 24 条
[1]  
Zhang L,Dong W,Zhang D,et al.Two-stage image denoisingby principal component analysis with local pixel grouping. Pattern Recognition . 2010
[2]  
Wang A,Liu L,Zeng B,et al.Progressive image coding basedon an adaptive block compressed sensing. IEICE ElectronicsExpress . 2011
[3]  
Candès E J,Romberg J,Tao T.Robust uncertainty principles:exact signal reconstruction from highly incomplete frequencyinformation. IEEE Transactions on Information Theory . 2006
[4]  
Deledalle C A,Salmon J,Dalalyan A.Image denoising withpatch based PCA:local versus global. Proceedings of the22nd British Machine Vision Conference . 2011
[5]  
Fowler J E,Mun S,Tramel E W.Multiscale blockcompressed sensing with smoother projected Landweberreconstruction. 19th European Signal ProcessingConference . 2011
[6]  
Do T,Gan L,Nguyen N,et al.Fast and efficient compressivesensing using structurally random matrices. IEEE Transactions on Electromagnetic Compatibility . 2012
[7]  
Candès E J,Romberg J.l1-magic:recovery of sparse signals viaconvex programming. http://www.acm.caltech.edu/l1magic . 2005
[8]  
Candès E,Wakin M.An introduction to compressive sam-pling——a sensing/sampling paradigm that goes against the common knowledge in data acquisition. IEEE Signal Processing Magazine . 2008
[9]  
Donoho D L.Compressed sensing. IEEE Transactions on In- formation Theory . 2006
[10]  
Gan L,Do T,Tran T D.Fast compressive imaging using scram-bled block Hadamard ensemble. Proceedings of the Euro-pean Signal Processing Conference . 2008