基于尺度空间理论和反扩散函数的图像去噪

Scale-Space and Inverse Diffusion Function Based Image Denoising

在使用扩散过程平滑噪声之后引入反扩散过程来恢复边缘,结合尺度空间理论和反扩散函数对图像进行去噪处理。该方法使用最小描述长度(M DL)准则自适应地选择图像中每一点处的最优尺度对图像进行滤波,加入尺度范围限制降低了过平滑和欠平滑的影响,改进了反扩散函数模型,对降质图像中的边缘进行恢复。与经典的滤波方法以及各向异性扩散方程的结果相比,本文方法取得了较好的效果。

An image denoising method is proposed by employing an inverse diffusion to restore edge information after image smoothing, combining the scale-space theory and the inverse diffusion function for image denoising. The novel method adopts the minimal description length （MDL） to self-adaptively choose the optimal scale for each pixel to smooth the image, and reduces the infection of over-smoothing or lack of smoothing by restricting the scale range. And it improves the inverse diffusion function to restore step edges on the degraded image. Experiments show that the method is more effective than classical filters and the anisotropic diffusion.