DocumentCode
2337346
Title
An automatic inspection system based on a neural network and uniform design
Author
Li, Meng-xin ; Wu, Cheng-dong ; Yue, Yong
Author_Institution
Shenyang Jianzhu Univ., China
Volume
7
fYear
2005
fDate
18-21 Aug. 2005
Firstpage
4528
Abstract
To solve the shortcomings of the traditional BP network, the improved algorithm is presented to accelerate the training and improve the accuracy, and reduce the possibility of getting into the local minimum. For optimal network structure, the UD method is introduced to optimise the parameters, and the ´best´ level-combination is obtained so that the performance of the classifier is further improved.
Keywords
automatic optical inspection; backpropagation; computer vision; image classification; neural nets; automatic inspection system; backpropagation; defect inspection; neural network; optimal network structure; parameter optimisation; pattern classification; uniform design; Acceleration; Algorithm design and analysis; Backpropagation algorithms; Design optimization; Electronic mail; Humans; Inspection; Machine learning; Neural networks; Production facilities; The improved BP algorithm; defect inspection; parameter optimization; uniform design;
fLanguage
English
Publisher
ieee
Conference_Titel
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location
Guangzhou, China
Print_ISBN
0-7803-9091-1
Type
conf
DOI
10.1109/ICMLC.2005.1527736
Filename
1527736
Link To Document