去除乘性噪声的重加权各向异性全变差模型

恢复含乘性噪声的图像是当前图像处理的重要研究课题. 本文提出基于迭代重加权的各向异性全变差(Total variation, TV)模型. 新模型中, 假定乘性噪声服从Gamma分布. 正则项采用加权的各向异性全变差, 其中, 自适应权函数由期望最大(Expectation maximization, EM)算法得到. 新模型在有效去噪的同时, 较好地保留了图像的边缘和细节信息, 同时能够有效地抑制"阶梯效应". 数值实验验证了新模型的效果.

Iteratively Reweighted Anisotropic-TV Based Multiplicative Noise Removal Model

Multiplicative noise removal is an important research topic on image processing. This paper proposes an iteratively reweighted anisotropic-total variation (TV) based model under the assumption that the multiplicative noise follows a Gamma distribution. The regularization term is the weighted anisotropic-TV regularizer. The weighting function incorporated in the regularization term is derived from the expectation maximization (EM) algorithm. The merits of this model are the preservation of geometrical structures of edges and the restraint of "staircase effect" while removing the noise. Numerical experimental results demonstrate the better performance of the proposed model.