多通道MCP范数最小化的彩色图像去噪方法  

Multi-channel MCPNorm Minimization for image denoising algorithm

作　　者：甘权森 雷雨 GAN Quansen;LEI Yun(School of Electronics and Information,Xi'an Polytechnic University,Xi'an 710048,China)

机构地区：[1]西安工程大学电子信息学院,陕西西安710048

出　　处：《长江信息通信》2022年第4期20-23,共4页Changjiang Information & Communications

摘　　要：低秩去噪方法利用核范数作为秩函数的凸近似,取得了很好的去噪效果。然而,非凸函数更加近似秩函数。针对彩色图像去噪问题,文章提出一种基于多通道MCP范数最小化的图像去噪方法。首先构建MCP(Minimax Concave Penalty)范数最小化的低秩去噪模型,对噪声图像的相似块矩阵施加MCP正则化约束。对于多通道且噪声强度不同的彩色图像,引入权重矩阵以平衡每个通道对恢复图像的估计。最后利用交替乘子方向法迭代求解模型,重建获得干净的图像。经实验验证,所提算法在性能指标和视觉效果方面都有较好的提升。The low-rank denoising method uses the nuclearnorm asa convex approximation of the rank function and achieves good denoising results.However,non-convex functions more closely the rank function.Based on the color image denoising problem,thispaper proposes an image denoising method based on Multi-channel Weighted MCP(Minimax Convave Penalty)norm minimization.Firstly,a MCP norm minimization based image denoising algorithm was constructed,andthe similar image patch matrix of the noise image was constraint by the non-convex MCP norm regularization.The multi-channel color image with different noise intensity was balancedby a weight matrixto the estimated restored image.Finally,a clean image is reconstructed by iterative solution of the model using alternate multiplier direction method.Experimental results show that the proposed algorithm can improve both performance and visual effect.

关 键 词：彩色图像去噪 MCP范数 非凸优化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]