摘要
图像压缩感知迭代重构算法主要采用迭代阈值法解决信号的重构问题,但是迭代阈值法仅仅利用变换系数进行阈值处理,并未考虑系数的邻域统计特性,导致重构性能不高。提出一种基于小波域滤波的迭代硬阈值迭代算法,利用小波域系数的邻域统计特性修订迭代硬阈值重构算法的代价函数,进行两步迭代收缩,并在迭代中用小波域滤波除去其中的重构噪声。实验结果表明,在相同的观测数据下,相比已有的经典算法,新算法的重构图像质量较高,并且可以获得快速的重构速度。
Iterative thresholding recovery method is used to solve the signal reconstruction problems in many iterative image compressed sensing techniques,which only treatment threshold itself and do not exploit the statistic characteristics of transform coefficients in neighborhood. A iterative hard thresholding recovery algorithm using wavelet filtering is presented,it improves the cost function with the statistical dependencies between transform coefficients,a two-step iterative shrinkage and wavelet filter which remove the reconstruction noise is used in iterative recovery. Experimental results show that,compared to other algorithms,the proposed algorithm can get higher image reconstruction performance,and get fast reconstruction speed with the same measurement rate.
出处
《电视技术》
北大核心
2014年第9期32-35,共4页
Video Engineering
关键词
信号处理
压缩感知
迭代硬阈值重构
小波域滤波
signal processing
compressed sensing
iterative hard threshold reconstruction
wavelet filtering
