An improved Average Filtering technique based on Statistical Trust Model

Image filtering is one of the most important fields of image enhancement, where unwanted intensity values, that corrupt the overall image quality is nullified to reproduce the original picture with minimum error of deviation. In this paper we proposed a method to improve the Classical Average Filtering technique based on Statistical Trust Model. The proposed method applies trust model to identify the unwanted intensities in the image, and replace those intensities with a dynamically calculated local mean value, thus ensuring better information retention and noise filtering simultaneously. The proposed scheme has been implemented on large number of corrupted digital images and has given encouraging results.