摘要
针对目前存在的压缩感知(CS)重建算法计算复杂度过高的问题,该文提出一种基于最佳线性估计的快速CS图像重建算法。该算法在编码端进行分块自适应CS随机测量,在解码端根据图像块不同的统计特性,估计出统计自相关函数矩阵,进而构造出最佳线性算子用于重建出各个图像块。由于该算法用线性投影的方式替代了传统CS重建算法的非线性迭代过程,使得其大大缩短了图像重建时间。仿真实验结果表明,对于纹理细节不复杂的图像,所提出的算法并没有因为其计算复杂度的减少而影响到重建质量,仍优于目前流行的CS重建算法。
Because of existing Compressed-Sensing (CS) reconstruction algorithms have high computing complexity a fast algorithm based on best linear estimate is proposed. It adaptively measures image data with a block-by-block manner at encoder, and reconstructs each block at decoder using the best linear operator which is constituted by statistical autocorrelation function matrix estimated according to various statistical property of image block. This algorithm replaces lots of nonlinear iterations in traditional CS reconstruction algorithm with linear projection, therefore it shorten the time of recovering image. Simulation experimental results indicate that the proposed algorithm not only reduces the time of rebuilding image, but also is better than the current popular CS reconstruction algorithm for images containing uncomplicated textures on the reconstructed image quality.
出处
《电子与信息学报》
EI
CSCD
北大核心
2012年第12期3006-3012,共7页
Journal of Electronics & Information Technology
基金
国家自然科学基金(61071091)
江苏高校优势学科"信息与通信工程"建设工程项目
江苏省普通高校研究生科研创新计划(CXZZ12_0466)
江苏省高校自然科学研究项目(12KJB510019)资助课题
关键词
图像处理
分块压缩感知
自适应测量
最佳线性估计
统计自相关函数
快速图像重建
Image processing
Block Compressed Sensing(BCS)
Adaptive measure
Best linear estimate
Statistical autocorrelation function
Fast image reconstruction