영상에서 임펄스 잡음제거를 위한 적응력 있는 가중 평균 필터

Abstract

In this work, a new adaptive weighted mean lter is proposed for preserving imagedetails while eectively suppressing impulse noise. The proposed lter is based ona noise pixel detection-estimation strategy. All the pixels are rst detected using animpulse noise detector. Then the detected noise pixels are replaced with the outputof the weighted mean lter over adaptive working window according to the rate ofcorrupted neighborhood pixels, while noise-free pixels are left unaltered.We compare the proposed lter to other existing lters in the qualitative measure andquantitative measures such as PSNR and MAE as well as computation time to verifythe capability of the proposed lter. Extensive simulations show that the proposed lterperforms better than other lters in impulse noise suppression and detail preservationwithout increasing of running time.