Presentation 2009-01-19
Handwritten Character Recognition by NG×SOM
Kouichi GUNYA, Makoto OTANI, Tetsuo FURUKAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We have proposed SOM^2 and NG×SOM algorithms. These algorithms have the ability of representing a set of data distributions. SOM^2 is the algorithm in which the reference vector units of the conventional SOM are replaced by the SOM modules. In the same way, NG×SOM has the SOM architecture with NG modules. In NG×SOM, each NG module represents a data distribution independently. As the result, NG×SOM sometimes fails to make an unappropriate interpolation between two data distributions. In this paper, we propose an novel algorithm in which the topological structure is unified though NG modules. The new NG×SOM was applied to hand written letter recognition, and the result shows the good performance of the algorithm.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) self-organizing maps / neural gas / SOM^2 / NG×SOM / Handwritten Character Recognition
Paper # NC2008-84
Date of Issue

Conference Information
Committee NC
Conference Date 2009/1/12(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) Handwritten Character Recognition by NG×SOM
Sub Title (in English)
Keyword(1) self-organizing maps
Keyword(2) neural gas
Keyword(3) SOM^2
Keyword(4) NG×SOM
Keyword(5) Handwritten Character Recognition
1st Author's Name Kouichi GUNYA
1st Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology()
2nd Author's Name Makoto OTANI
2nd Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology
3rd Author's Name Tetsuo FURUKAWA
3rd Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology
Date 2009-01-19
Paper # NC2008-84
Volume (vol) vol.108
Number (no) 383
Page pp.pp.-
#Pages 6
Date of Issue