A Novel Iterative and Adaptive Noise Detection Method for Removal of Impulse Noise

摘要

In this paper, the authors propose an efficient technique called Iterative and Adaptive Noise Detection Method (IANDM) for detecting the corrupted pixel in the image and for estimating the original value of the corrupted pixel. With the proposed method the noise pixel detection is improved by giving important to both the local and the global features. The proposed AMPMAD algorithm has a highest detection in terms of 99.99% for 15% corrupted pixels in Lena image. Further the authors propose an adaptive iterative mode replacement policy for estimating the values of corrupted pixels. The performance evaluation of the proposed noise detector is measured viz PFN and PFP. The quantitative evaluation of the filter measured in terms of PSNR and MSE.