Presentation 1998/2/19
A Discriminant Function for Noisy Pattern Recognition and Its Application to Blurred Chinese Character Recognition
Shin'ichiro OMACHI, Fang SUN, Hirotomo ASO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, a new discriminant function for recognition of noisy patterns is presented. If noise is added to input data, the distribution of feature vectors will change according to the noise. In this case, probability density function of distribution is considered to be necessary to reflect the changes. We investigate how distribution of vectors changes if noise is added to some elements of vectors, and a discriminant function that reflects the changes is proposed. This method supplements the existing statistical methods and has high ability of recognition of data with noise. The effectiveness of the new method is shown by simulations with artificial data and experiments with real blurred character images.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) character recognition / Mahalanobis distance / Bayes classifier / distribution
Paper #
Date of Issue

Conference Information
Committee PRMU
Conference Date 1998/2/19(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 Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Discriminant Function for Noisy Pattern Recognition and Its Application to Blurred Chinese Character Recognition
Sub Title (in English)
Keyword(1) character recognition
Keyword(2) Mahalanobis distance
Keyword(3) Bayes classifier
Keyword(4) distribution
1st Author's Name Shin'ichiro OMACHI
1st Author's Affiliation School of Engineering, Tohoku University()
2nd Author's Name Fang SUN
2nd Author's Affiliation School of Engineering, Tohoku University
3rd Author's Name Hirotomo ASO
3rd Author's Affiliation School of Engineering, Tohoku University
Date 1998/2/19
Paper #
Volume (vol) vol.97
Number (no) 558
Page pp.pp.-
#Pages 8
Date of Issue