Title :
Multilayer neural networks in handwritten recognition
Author :
Ciota, Zygmunt ; Gawinowski, Marcin
Author_Institution :
Dept. of Microelectron. & Comput. Sci., Tech. Univ. Lodz, Poland
Abstract :
Multilayer neural network presented in This work and it can be helpful for recognition of handwritten words. Learning efficiency of such network also has been presented. For high learning efficiency a standard backpropagation method has been extended by adding a momentum term. Additional analysis shows how to improve proposed solutions using mixed hardware-software co-design and fuzzy logic approach.
Keywords :
backpropagation; feedforward neural nets; fuzzy logic; handwriting recognition; hardware-software codesign; learning (artificial intelligence); multilayer perceptrons; optical character recognition; backpropagation method; feedforward neural network; fuzzy logic; handwritten recognition; learning; mixed hardware software codesign; multilayer neural networks; Artificial neural networks; Backpropagation algorithms; Character recognition; Handwriting recognition; Image recognition; Intelligent networks; Multi-layer neural network; Neural networks; Neurons; Pattern recognition;
Conference_Titel :
Modern Problems of Radio Engineering, Telecommunications and Computer Science, 2004. Proceedings of the International Conference
Conference_Location :
Lviv-Slavsko, Ukraine
Print_ISBN :
966-553-380-0
