Title :
Design of weighted order statistic filters by training-based optimization
Author :
Koivisto, Pertti ; Huttunen, Heikki
Author_Institution :
Signal Process. Lab., Tampere Univ. of Technol., Finland
Abstract :
This paper demonstrates how weighted order statistic filters can be designed using training-based optimization. The design method utilizes supervised learning with simulated annealing as the learning rule. In addition, the efficiency and flexibility of the presented method are studied through experiments
Keywords :
circuit optimisation; image processing; learning (artificial intelligence); median filters; network synthesis; nonlinear filters; simulated annealing; design method efficiency; heavy-tailed noise; impulsive noise; learning rule; noise distribution; nonlinear filters; simulated annealing; supervised learning; training image; training-based optimization; weighted median filters; weighted order statistic filter design; Design methodology; Design optimization; Filters; Laboratories; Optimization methods; Signal design; Signal processing; Simulated annealing; Statistics; Supervised learning;
Conference_Titel :
Signal Processing and its Applications, Sixth International, Symposium on. 2001
Conference_Location :
Kuala Lumpur
Print_ISBN :
0-7803-6703-0
DOI :
10.1109/ISSPA.2001.949770
