Presentation	1998/6/19 Handwritten Kanji Recognition Using Difference Principal Components in the Complementary Classifier Takahiko KAWATANI,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	When a complementary classifier with an existent one is designed to improve recognition accuracy, it is important to obtain features which make it easy to recognize patterns misrecognized in the existent classifier. This paper proposes to extract difference principal components for each class which are effective to separate patterns belonging to each class from those misrecognized in the class. When the complementary classifer was combined with the classifier designed using LDA method, it was confirmed that the recognition accuracy for test data increases from 97.35% to 98.00% by using five difference principal components extracted from common difference principal components.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	character recognition / classifier combination / difference principal component / LDA / handwritten Kanji
Paper #	PRMU98-39
Date of Issue

Conference Information
Committee	PRMU
Conference Date	1998/6/19(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)	Handwritten Kanji Recognition Using Difference Principal Components in the Complementary Classifier
Sub Title (in English)
Keyword(1)	character recognition
Keyword(2)	classifier combination
Keyword(3)	difference principal component
Keyword(4)	LDA
Keyword(5)	handwritten Kanji
1st Author's Name	Takahiko KAWATANI
1st Author's Affiliation	Hewlett-Packard Laboratories Japan()
Date	1998/6/19
Paper #	PRMU98-39
Volume (vol)	vol.98
Number (no)	127
Page	pp.pp.-
#Pages	8
Date of Issue