An Anisotropic Wiener Denoising Algorithm Based on Structure Tensor
结合结构张量和Wiener滤波的图像去噪算法

An improved Wiener denoising algorithm is proposed by combining structure tensor of image together with Wiener filtering, which is data-adaptive and thus can better preserve boundaries and reduce structure deloealization. An additive operator splitting scheme is applied in solving the nonlinear diffusion equation to improve computation efficiency. The experimental results show that the peak signal-to-noise ratio of the denoised image is improved via the proposed algorithm, as well as the visual effect.