Paper Info
Title
Evolutionary tree-structured filter for impulse noise removal
Abstract
A new evolutionary approach for construction of uniform impulse noise filter is presented. Genetic programming is used for combining the basic image transformations and filters into tree structure, which can accurately estimate noise map. Proposed detector is employed for building switching-scheme filter, where recursively implemented α-trimmed mean is used as the estimator of corrupted pixel values. The proposed evolutionary filtering structure shows very good results in removal of uniform impulse noise, for wide range of noise probabilities and different test images.
Year
DOI
Venue
2006
10.1007/11864349_10
ACIVS
Keywords
Field
DocType
new evolutionary approach,uniform impulse noise filter,tree structure,uniform impulse noise,evolutionary tree-structured filter,proposed detector,basic image transformation,noise probability,noise map,impulse noise removal,corrupted pixel value,different test image,evolutionary trees,impulse noise
Value noise,Median filter,Digital filter,Pattern recognition,Noise measurement,Computer science,Salt-and-pepper noise,Algorithm,Impulse noise,Artificial intelligence,Gaussian noise,Gradient noise
Conference
Volume
ISSN
ISBN
4179
0302-9743
3-540-44630-3
Citations 
PageRank 
References 
4
0.45
7
Authors
2
Name
Order
Citations
PageRank
Nemanja I. Petrović180.94
Vladimir S. Crnojevic218617.82