High density impulse noise removal by Fuzzy Mean Linear Aliasing Window Kernel

• Published in: 2012 IEEE International Conference on Signal Processing, Communication and Computing (ICSPCC) pp. 711 - 716
• Main Authors: Utaminingrum, Fitri, Uchimura, K., Koutaki, G.
• Format: Conference Proceeding
• Language: English; Japanese
• Published: IEEE 01.08.2012
• Subjects: Filtering theory; fuzzy method; impulse noise removal; Kernel; linear aliasing filter; Maximum likelihood detection; Nonlinear filters; PSNR
• ISBN: 9781467321921; 1467321923
• DOI: 10.1109/ICSPCC.2012.6335693

Fuzzy Mean Linear Aliasing Window Kernel (FMLAWK) filter method proposed to reducing the high-density impulse noise interference and generating the smooth image performance. FMLAWK filter is a spatial filter, which combined from fuzzy method and Linear Aliasing Filter (LAF). The initial step is finding the degree of membership function (μ) value of each matrix element on the corrupted image which use the fuzzy method. Furthermore, the μ value of the corrupted image processed by LAF method which using 3×3 window. The reducing of 3×3 windows on LAF process will be obtain one pixel data based on Linear method. Our research also provides kernel algorithms. Preprocessing Kernel algorithm used for checking of each element matrix on the 3×3 window. If the matrix element contaminated by impulse noise, so the matrix element replaced with a new element data. Our simulation result shows the image filtering better and smoother quality than the comparison method.