Title :
Pattern recognition using GAMLP systems: A comparative view as performance level
Author :
Iulian-Constantin Vizitiu;Petrică Ciotîrnae;Andrei Ko Vacs;Mădălina Mazîlu
Author_Institution :
Military Technical Academy, Communications and Electronic Systems Department, Bucharest, Romania
fDate :
6/1/2012 12:00:00 AM
Abstract :
According to the special literature, a major research direction to improve the performance level assigned to a pattern recognition system is to use the specific paradigms of artificial intelligence domain. Having as starting point the indubitable advantages given by GANN (Genetic Algorithm Neural Nework) system concept in solving of some concrete pattern recognition tasks, in this paper a comparative view as performance level between two types of GAMLP (Genetic Algorithm MLP neural network) systems is described. Finally, using a real database belonging to high-resolution radar imagery, the obtained results confirm the great potential of the designed GAMLP systems to be integrated inside of real pattern recognition applications.
Keywords :
"Genetics","Pattern recognition","Biological cells","Genetic algorithms","Neural networks","Training","Standards"
Conference_Titel :
Communications (COMM), 2012 9th International Conference on
Print_ISBN :
978-1-4577-0057-6
DOI :
10.1109/ICComm.2012.6262578