DocumentCode
2665586
Title
A new feature extraction method by neural networks
Author
Jou, I-Chang ; Yu, Shih-Shien ; Tsay, S.-C.
Author_Institution
Minist. of Commun., Chung-Li, Taiwan
fYear
1990
fDate
1-3 May 1990
Firstpage
3249
Abstract
A neural network which is composed from the back propagation (BP) network and the neocognitron network is proposed. This neural network has the capability of high feature extraction for pattern recognition. Experimental results show that the recognition rate is 92% for the recognition of handwritten digits, which is better than the one hidden-layer BP method by about 14%
Keywords
neural nets; pattern recognition; back propagation; feature extraction; handwritten digits; neocognitron network; neural networks; pattern recognition; Backpropagation; Data mining; Feature extraction; Image recognition; Neural networks; Neurons; Parallel processing; Pattern recognition; Transfer functions;
fLanguage
English
Publisher
ieee
Conference_Titel
Circuits and Systems, 1990., IEEE International Symposium on
Conference_Location
New Orleans, LA
Type
conf
DOI
10.1109/ISCAS.1990.112704
Filename
112704
Link To Document