自适应非局部低秩的图像压缩感知重构算法

Adaptive nonlocal low-rank based image compressed sensing reconstruction

针对传统非局部低秩的图像压缩感知重构算法忽略图像结构特征,导致图像重构效果不理想的问题,提出一种自适应非局部低秩的图像压缩感知重构算法,充分考虑图像自身结构特征和图像块间的强相关性。根据样本块的块结构稀疏度值设置阈值,自适应选取局部搜索窗口大小和相似块的数目;利用新的相似块匹配方法在给定搜索窗口内选取所需要的相似块,按列聚合成低秩矩阵;利用加权Schatten p-范数作为原始秩函数的逼近去求解矩阵秩最优化问题。实验结果表明,所提算法较对比算法在峰值信噪比和视觉效果上均有所提高,验证了其有效性。

To solve the problem that the traditional compressed sensing algorithm based on nonlocal low-rank model ignores the characteristics of image structure,which leads to unideal effects of image reconstruction,an image compressed sensing reconstruction algorithm based on adaptive nonlocal low-rank model was proposed.The strong relationship between the image patches and the structural features of the image itself was considered.According to the patch structured sparsity value of the sample image patches,the threshold was set,and the size of the local search window and the number of similar patches were selected,adaptively.The required similar patches were selected in the specified search window using the new similarity patch matching method,and the low-rank matrix was synthesized according to the column aggregation.A weighted Schatten p-norm was used as an approximation of the original rank function to solve the matrix rank optimization problem.Experimental results show that,compared with the contrast algorithm,the proposed algorithm improves the peak signal to noise ratio and visual perception,which verifies the effectiveness of the algorithm.