AN AMELIORATED DETECTION STATISTICS FOR ADAPTIVE MASK MEDIAN FILTRATION OF HEAVILY NOISED DIGITAL IMAGES

Abstract
Noise reduction is an important area of research in image processing applications. The performance of the digital image noise filtering method primarily depends upon the accuracy of noise detection scheme. This paper presents an effective detector based, adaptive mask, median filtration of heavily noised digital images affected with fixed value (or salt and pepper) impulse noise. The proposed filter presents a novel approach; an ameliorated Rank Ordered Absolute Deviation (ROAD) statistics to judge whether the input pixel is noised or noise free. If a pixel is detected as corrupted, it is subjected to adaptive mask median filtration; otherwise, it is kept unchanged. Extensive experimental results and comparative performance evaluations demonstrate that the proposed filter outperforms the existing decision type, median based filters with powerful noise detectors in terms of objective performance measures and visual retrieviation accuracy.

Authors
Geeta Hanji1, M.V. Latte2, K. Varsha3
P.D.A. College of Engineering, India1,3, JSS Academy of Technical Education, India2

Keywords
Rank Ordered Absolute Deviation (ROAD), Adaptive Mask, Noise Detector, Edge Preservation, Visual Retrieviation Accuracy
Published By :
ICTACT
Published In :
ICTACT Journal on Image and Video Processing
( Volume: 6 , Issue: 2 )
Date of Publication :
November 2015

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.