Presentation 2000/2/22
Holistic Word Recognition Based on Composition of Character Features
Yoshinobu Hotta, Hiroaki Takebe, Satoshi Naoi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Holistic word matching is effective for handwritten address recognition, because it reduces segmentation errors. But the fact that its dictionary size is proportional to the number of words, occurs capacity problem. In this paper, we propose a new word recognition method in which holistic word matching is executed without word feature dictionary. In our method, character feature dictionary is only needed and word feature dictionary is composed of character features. Experimental results show that proposed method can drastically reduce dictionary size although recognition rate drops from 90.6% to 87.4% comparing to conventional method which utilize word feature dictionary by extracting word features from word images.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Handwritten words / Feature composition / Word recognition / Multi-template
Paper # PRMU99-239
Date of Issue

Conference Information
Committee PRMU
Conference Date 2000/2/22(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) Holistic Word Recognition Based on Composition of Character Features
Sub Title (in English)
Keyword(1) Handwritten words
Keyword(2) Feature composition
Keyword(3) Word recognition
Keyword(4) Multi-template
1st Author's Name Yoshinobu Hotta
1st Author's Affiliation FUJITSU LABORATORIES LTD.()
2nd Author's Name Hiroaki Takebe
2nd Author's Affiliation FUJITSU LABORATORIES LTD.
3rd Author's Name Satoshi Naoi
3rd Author's Affiliation FUJITSU LABORATORIES LTD.
Date 2000/2/22
Paper # PRMU99-239
Volume (vol) vol.99
Number (no) 649
Page pp.pp.-
#Pages 8
Date of Issue