DocumentCode :
2031069
Title :
Design of multistage weighted order statistic filters by a neural network
Author :
Pi, Yiming ; Huang, Shunji
Author_Institution :
Beijing Univ. of Posts & Telecommun., China
Volume :
1
fYear :
1995
fDate :
23-26 Oct 1995
Firstpage :
346
Abstract :
A genetic back-propagation (BP) algorithm is used to solve the optimal design problem of multistage weighted order statistic filter under the mean absolute error criterion. It is shown experimentally that this algorithm can find the optimal WOS filter in image restoration application
Keywords :
adaptive filters; adaptive signal processing; backpropagation; circuit optimisation; digital filters; image restoration; multilayer perceptrons; nonlinear filters; adaptive filter; genetic backpropagation algorithm; image restoration; mean absolute error criterion; multilayer neural network; multistage weighted order statistic filters; nonlinear digital filter; optimal WOS filter; optimal design; Adaptive filters; Algorithm design and analysis; Boolean functions; Digital filters; Error analysis; Genetics; Multi-layer neural network; Neural networks; Stacking; Statistics;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Image Processing, 1995. Proceedings., International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-8186-7310-9
Type :
conf
DOI :
10.1109/ICIP.1995.529717
Filename :
529717
Link To Document :
بازگشت