Presentation 1993/12/14
Improved Vector Quantization and Its Application to Character Recognition
Atsushi Satou, Jun Tsukumo,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this report,the criterion of training reference vectors is formulalted on the analogy of the back propagation algorithm for multi layer perceptron(MLP),which decreases the error rate for samples dose to dedsion boundaries.We present improved vector quantization method based on the aboveidea,and decsion boundaries by this method are discussed compared with LVQ2.Experimental resultes for printed Japanese Hiragana characters recognition reveal that the proposed method is superior to LVQ2 and MLP.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Neural Network / Learning Vector Quantization / Template Matching / Character Recognition / Pattern Recognition
Paper # NC93-60
Date of Issue

Conference Information
Committee NC
Conference Date 1993/12/14(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) Improved Vector Quantization and Its Application to Character Recognition
Sub Title (in English)
Keyword(1) Neural Network
Keyword(2) Learning Vector Quantization
Keyword(3) Template Matching
Keyword(4) Character Recognition
Keyword(5) Pattern Recognition
1st Author's Name Atsushi Satou
1st Author's Affiliation Information Technology Research Laboratories,NEC()
2nd Author's Name Jun Tsukumo
2nd Author's Affiliation Information Technology Research Laboratories,NEC
Date 1993/12/14
Paper # NC93-60
Volume (vol) vol.93
Number (no) 376
Page pp.pp.-
#Pages 9
Date of Issue