Title :
A novel hybrid filter for edge preserving restoration of image
Author :
Qian, Xiaoliang ; Guo, Lei ; Yu, Bo ; Zhao, Tianyun
Author_Institution :
Coll. of Autom., Northwestern Polytech. Univ., Xi´´an, China
Abstract :
In order to realize the edge preserving restoration of image, a novel hybrid filter which is a combination of Tukey anisotropic diffusion and an adaptive median filter based on object scale is proposed in this paper. The Tukey anisotropic diffusion can suppress the most of noises while preserving the edges, however, the impulse noise points which usually have large gradient will be regarded as edges by Tukey anisotropic diffusion and also be preserved. Our proposed adaptive median filter identify the impulse noise point adaptively according to the object scale of pixel, thus, it can not only remove the impulse noise clearly but also avoid the filtering of the most of pixels. The simulation which add the Gaussian noises and pepper-and-salt noises to Lena indicate that the ability of our method in preserving the fine details is superior than several traditional image filter.
Keywords :
Gaussian noise; adaptive filters; image restoration; impulse noise; median filters; Gaussian noises; Tukey anisotropic diffusion; adaptive median filter; edge preserving restoration; hybrid filter; image restoration; impulse noise; object scale; pepper-and-salt noises; Filtering algorithms; Image edge detection; Variable speed drives; Tukey; adaptive median filter; anisotropic diffusion; impulse noise; object scale;
Conference_Titel :
Advanced Computer Theory and Engineering (ICACTE), 2010 3rd International Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4244-6539-2
DOI :
10.1109/ICACTE.2010.5579690