基于非局部相似块低秩的压缩感知图像重建算法

发布时间：2018-03-11 17:22

本文选题：压缩感知　切入点：图像重建　出处：《电子学报》2017年03期 　论文类型：期刊论文

【摘要】：传统的压缩感知重建算法利用信号在某个特征空间下的稀疏性构建目标优化函数,但没有充分考虑信号的局部特性和结构化属性,影响了算法的重建性能和算法的适应性.本文考虑图像的非局部自相似性(Nonlocal Self-Similarity,NLSS),提出一种基于图像相似块低秩的压缩感知图像重建算法,将图像恢复问题转化为聚合的相似块矩阵秩最小问题.算法以最小压缩感知重建误差为约束构建优化模型,并采用加权核范数最小化算法(Weighed Nuclear Norm Minimization,WNNM)求解低秩优化问题,很好地挖掘了图像自身的信息和结构化稀疏特征,保护了图像的结构和纹理细节.多个测试图像、不同采样率下的实验证明了算法的有效性,特别是在低采率下对于纹理较为丰富的图像,提出的算法图像重建质量较明显的优于最新的同类算法.
[Abstract]:The traditional compressed perceptual reconstruction algorithm uses the sparsity of the signal in a feature space to construct the objective optimization function, but it does not fully consider the local and structural properties of the signal. The performance of the algorithm and the adaptability of the algorithm are affected. In this paper, considering the nonlocal self-SimilityNLSSs of non-local self-similarity of images, a compressed perceptual image reconstruction algorithm based on image similarity block and low rank is proposed. The image restoration problem is transformed into the rank minimization problem of aggregated similar block matrices. The algorithm uses minimum compression perceptual reconstruction error as the constraint to construct the optimization model, and uses weighted Nuclear Norm minimization algorithm (WNNM) to solve the low rank optimization problem. The structure and texture details of the image are well mined, and the structure and texture details of the image are protected. The experimental results of several test images and different sampling rates demonstrate the effectiveness of the algorithm. Especially for the images with abundant texture, the reconstruction quality of the proposed algorithm is obviously better than that of the latest similar algorithms.
【作者单位】：长沙理工大学综合交通运输大数据智能处理湖南省重点实验室;长沙理工大学计算机与通信工程学院;湖南大学信息科学与工程学院;中国科学院计算技术研究所;
【基金】：国家自然科学基金(No.61471343,No.61572183,No.61402053) 湖南省教育厅科学研究重点项目(No.13A107,No.15A007) 湖南省自然科学基金(No.2016JJ2005) 湖南省科技计划项目(No.2014FJ6047,No.2014GK3030)
【分类号】：TP391.41

【相似文献】