Estimation of Optimal Parameter in ε-Filter Based on Signal-Noise Decorrelation

Mitsuharu MATSUMOTO  Shuji HASHIMOTO 

Publication
IEICE TRANSACTIONS on Information and Systems  Vol.E92-D  No.6  pp.1312-1315
Publication Date: 2009/06/01
Online ISSN: 1745-1361
Print ISSN: 0916-8532
Type of Manuscript: LETTER
Category: Algorithm Theory
Keyword: 
parameter optimization,  nonlinear filter,  ε-filter, 

Full Text: PDF(284.5KB)

Summary: 
ε-filter is a nonlinear filter for reducing noise and is applicable not only to speech signals but also to image signals. The filter design is simple and it can effectively reduce noise with an adequate filter parameter. This paper presents a method for estimating the optimal filter parameter of ε-filter based on signal-noise decorrelation and shows that it yields the optimal filter parameter concerning a wide range of noise levels. The proposed method is applicable where the noise to be removed is uncorrelated with signal, and it does not require any other knowledge such as noise variance and training data.