Article details

Title: More Efficient ATR System Using a New Neural Classification Chain
Author(s): Constantin-Iulian Vizitiu   Stelian Spînu   Grigore Jeler         

Abstract: One of the most important methods to improve the quality of the recognition function assigned to modern automatic target recognition (ATR) systems is to use, in a proper way, new powerful artificial neural architectures in the classification chain. Consequently, an improved neural classification chain using a new procedure for pattern feature extraction, a generalized version of Sammon projection algorithm, and a genetic algorithm to increase the performances of RBF neural networks are proposed. To confirm the broached theoretical aspects, a real image database is used.

Keywords: ATR system, feature extraction and selection, genetic algorithm, RBF neural network.

References:
[1] F. SADJADI, B. JAVIDI (Eds.) – Physics of Automatic Target Recognition, Springer, New York, NY, 2007
[2] G. O’BRIEN – Automatic Target Recognition, In R.G. Driggers (Ed.), “Encyclopedia of Optical Engineering”, Vol. 1, pp. 114-120, Marcel Dekker, New York, NY, 2003
[3] C.I. VIZITIU, P. CIOTÎRNAE – Neural Design Procedure for an ATTR System Based on Video Imaginary Usage, Proc. of the 9th WSEAS International Conference on Automation and Information, ICAI ’08, pp. 525-530, Bucharest, Romania, Jun. 24-26, 2008
[4] M.W. ROTH – Survey of Neural Network Technology for Automatic Target Recognition, IEEE Transactions on Neural Networks, Vol. 1, No. 1, pp. 28-43, Mar. 1990
[5] C.I. VIZITIU – Neural Networks Used in Video Pattern Recognition, Military Technical Academy Publishing House, Bucharest, Romania, 2002 (in Romanian)
[6] C.I. VIZITIU – Genetic Algorithms and Neural Networks: Theory and Applications, Military Technical Academy Publishing House, Bucharest, Romania, 2004 (in Romanian)
[7] R.O. DUDA, P.E. HART, D.G. STORK – Pattern Classification, 2nd Edition, John Wiley & Sons, New York, NY, 2001
[8] C.I. VIZITIU, D. MUNTEANU, C. MOLDER – A New Version of Flusser Moment Set for Pattern Feature Extraction, WSEAS Transactions on Information Science and Applications, Vol. 5, No. 4, pp. 396-406, Apr. 2008
[9] C.I. VIZITIU, D. MUNTEANU – A Genetic Procedure for RBF Neural Networks Centers Selection, Proc. of the 9th WSEAS International Conference on Neural Networks, NN ’08, pp. 37-40, Sofia, Bulgaria, May 2-4, 2008