基于结构相似度的图像去噪新方法

A Novel Image Denoising Method Based on Structural Similarity

将结构相似度作为一种刻画忠诚项的度量用于图像去噪模型中。针对经典ROF模型忠诚项的约束项L2度量未考虑图像空间结构性而导致恢复图像视觉效果差的缺陷,引入结构相似度来改进模型的忠诚项,提出了一种新的去噪模型。为在去噪过程中,更好地保护图像的边缘,在此模型的基础上,文中还做了进一步改进,用非凸正则项代替TV正则项,得到推广模型。实验结果表明,相对于ROF模型,两个模型在有效去除噪声的同时,能更好地保持图像的结构信息,提高图像的视觉效果,且推广模型在图像边缘保护方面的性能更好。

In this paper, we make one of the first attempts to incorporate the structural similarity as the fidelityterm into the framework of image denoising in order to overcome the weakness that the fidelity term of the classicalROF model does not consider image structure, which leads to poor visual image restoration. This paper first proposesa new image denoising model （ model 1 ） by introducing structural similarity as the fidelity term instead of the origi-nal. In promotion model 2, a nonconvex regularization rather than the classical TV regularization is used to preservethe edge better while removing noises. Experimental results show model 1 and model 2 achieve better performancethan ROF model in image structural information and perceptual image quality. And model 2 plays a more importantrole in preserving image edge than model 1.