Adaptive Right Median Filter for Salt-and-Pepper Noise Removal

Authors : Uğur ERKAN, Levent GÖKREM, Serdar ENGİNOĞLU

Pages : 542-550

Doi:10.29137/umagd.495904

View : 18 | Download : 6

Publication Date : 2019-06-30

Article Type : Research Paper

Abstract :In image processing, nonlinear filters are commonly used as a pre-process for noise removal before applying any advanced processing such as classification and clustering to an image. The adaptive filters being a kind of the nonlinear filters mainly perform better than the others in salt-and-pepper noise. In this paper, we first define a new median method, i.e. right median insert ignore into journalissuearticles values(rm);. We then define a new adaptive nonlinear filter developed via rm, namely Adaptive Right Median Filter insert ignore into journalissuearticles values(ARMF);, for saltand-pepper noise removal. Afterwards, we compare the results of ARMF with some of the known filters by using 12 test images and two image quality metrics: Peak Signal-to-Noise Ratio insert ignore into journalissuearticles values(PSNR); and Structural Similarity insert ignore into journalissuearticles values(SSIM);. The results show that ARMF outperforms the other methods at all the noise density except 80% and 90% in the mean percentages. Finally, we discuss the need for further research.
Keywords : Image denoising, Noise removal, Nonlinear filters