DocumentCode :
2206398
Title :
An image recognition method based on multiple BP neural networks fusion
Author :
Ni, Jianjun ; Ma, Xiaoping ; Xu, Lizhong ; Wang, Jianying
Author_Institution :
Inf. & Electr. Eng. Coll., China Univ. of Mining & Technol., Xuzhou, China
fYear :
2004
fDate :
21-25 June 2004
Firstpage :
323
Lastpage :
326
Abstract :
With respect to the limitation of the image recognition method by single BP neural network, a new method based on multiple BP neural networks with D-S theory is proposed. By using this method to synthesize each image recognition result of single BP neural network, we can improve the performance of the image recognition. The results of the computer simulation and experiment show that the new method is effective.
Keywords :
backpropagation; belief maintenance; case-based reasoning; image recognition; neural nets; probability; sensor fusion; D-S evidence theory; belief function; computer simulation; image recognition method; multiple BP neural networks fusion; probability assignment; single BP neural network; Computer networks; Computer simulation; Educational institutions; Feedforward neural networks; Image recognition; Image resolution; Multi-layer neural network; Network synthesis; Neural networks; Pattern recognition;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Acquisition, 2004. Proceedings. International Conference on
Print_ISBN :
0-7803-8629-9
Type :
conf
DOI :
10.1109/ICIA.2004.1373380
Filename :
1373380
Link To Document :
بازگشت