Skip to Main Content
To process an image is to treat the images properly, such as in the applications of image noise reduction. The goal of noise reduction is to reduce the damage to the image detail and blur. Many researches are to determine the content of images from pixels. Some researches focus on reducing the noise by the image classification and estimation. The thresholds of these classification algorithms are always depending on their experimental results, or on the fixed form of filtering method. This paper presents a method of classification and estimation for impulse noise to reduce the damage to the image detail and blur. Firstly, we get the standard deviation as the threshold for each pixel in the image. After that, we compare the center pixel to the neighbor pixels of four directions with vertical and horizontal extent. If the difference is greater than the threshold, we judge the center pixel as a noise. The experimental results show that the proposed algorithm can reduce the erroneous judgment and enhance the performance of image restoration.