An effective adaptive technique for impulse noise detection and reduction in digital images

In this paper, an effective method proposed for detects and removal of impulse noise from digital images. In our method consists of two stages: in first stage we detect noisy pixel based on intensity values. In the second stage we replace noisy pixel with average estimated from noise-free pixels within a small neighborhood. The method adaptively changes the size of the noise-free pixel in the neighborhood. The results show this method is good for noise reduction and image restoration especially in high level noisy images.