Using Hierarchical Adaptive Neuro Fuzzy Systems And Design Two New Edge Detectors In Noisy Images
Mohammad Hossein Olyaee, H. Abasi, Mahdi Yaghoobi · Journal of Soft Computing and Applications · 2013
Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. One of the most important topics in image processing is edge detection. Many methods have been proposed for this end but most of them have weak performance in noisy images because noise pixels are determined as edge. In this paper, two new methods are represented based on Hierarchical Adaptive Neuro Fuzzy Systems (HANFIS). Each method consists of desired number of HANFIS operators that receive the value of some neighbouring pixels and decide central pixel is edge or not. Simple train images are used in order to set internal parameters of each HANFIS operator. The presented methods are evaluated by some test images and compared with several popular edge detectors. The experimental results show that these methods are robust against impulse noise and extract edge pixels exactly.