Presentation 1994/2/28
Improvement of Selective Attention Model
Kenji Inada, Kunihiko Fukushima,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) ″Selective attention model″proposed by Fukushima has a large ab ility to recognize composite figures consisting of many patterns. Previously,Ashida et al.has attempted to use this model to recognize Chinese Characters,but the ability of their system to recognize distorted patterns was not large enough.We propose an improved Chinese character recognition system using the mechanism of the selective attention model.In our model,the strength of inhibitory connections on backward path is set larger than that of the forward path.On the other hand,Tanigawa et al.has proposed a neocognitoron which has threshold-control mechanism with local- feedback to improve recognition rate.Therefore,we introduced the threshold-control mechanism into the selective attention model.We have confirmed by computer simulation that our system recognizes Chinese characters more robustly than the previous model.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) selective attension model / local-feedback / threshold-control / Chinese character recognition
Paper # NC93-78
Date of Issue

Conference Information
Committee NC
Conference Date 1994/2/28(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Improvement of Selective Attention Model
Sub Title (in English)
Keyword(1) selective attension model
Keyword(2) local-feedback
Keyword(3) threshold-control
Keyword(4) Chinese character recognition
1st Author's Name Kenji Inada
1st Author's Affiliation Department of Biophysical Engineering,Faculty of Engineering Science,Osaka University()
2nd Author's Name Kunihiko Fukushima
2nd Author's Affiliation Department of Biophysical Engineering,Faculty of Engineering Science,Osaka University
Date 1994/2/28
Paper # NC93-78
Volume (vol) vol.93
Number (no) 490
Page pp.pp.-
#Pages 8
Date of Issue