用于图像平滑去噪的快速各向异性反扩散算法

提出一种用于灰度图像平滑去噪的基于自适应统计滤波的快速各向异性反扩散算法.该算法通过建立一种改进的异性扩散算法,对含噪图像迭代地进行反扩散作用,并在两次反扩散之间实行自适应统计滤波,以达到平滑去噪的效果.实验结果表明,新算法的性能稳定,相对于已有算法能在更短时间内达到预期效果,同时使图像边缘得到一定增强.

Fast Anisotropic Inverse Diffusion Based on Adaptive Statistic Filtering

A fast anisotropie inverse diffusion algorithm based on adaptive statistic filtering （FAID-ASF）is proposed in this paper. A reformed anisotropie diffusion model is used for inverse diffusion of noising images. Between two steps of inverse diffusion, adaptive statistic filtering is performed so that the high noises left by the inverse diffusion are eliminated. According to the experiment results, the new algorithm has a steady performance. Moreover, it achieves a better and efficient ability of smoothing, de-noising and edge preserving compared with the known methods.