Presentation 2013-12-12
Character Recognition using Statistical Evaluation of Surrounding Information
Naoki UCHIDA, Yuji WAIZUMI, Kazuyuki TANAKA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Scene text recognition methods often consist of text extraction step and recognition step. Recognition accuracy of this two step approach can degrade due to the cumulated errors in the two steps. In this report, we propose a scene text recognition method which evaluates shapes of regions in a scene image and relative locations among them simultaneously, can recognize texts without text extraction process. The proposed method evaluates the distributions of two features for shapes and relative locations with Gaussian distributions, and recognize texts in an image using a score based on likelihood density.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Text Recognition / Statistical Evaluation / Positional Relation between Characters / Gaussian Distribution
Paper # PRMU2013-70
Date of Issue

Conference Information
Committee PRMU
Conference Date 2013/12/5(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) Character Recognition using Statistical Evaluation of Surrounding Information
Sub Title (in English)
Keyword(1) Text Recognition
Keyword(2) Statistical Evaluation
Keyword(3) Positional Relation between Characters
Keyword(4) Gaussian Distribution
1st Author's Name Naoki UCHIDA
1st Author's Affiliation Graduate School of Information Sciences, Tohoku University()
2nd Author's Name Yuji WAIZUMI
2nd Author's Affiliation Graduate School of Information Sciences, Tohoku University
3rd Author's Name Kazuyuki TANAKA
3rd Author's Affiliation Graduate School of Information Sciences, Tohoku University
Date 2013-12-12
Paper # PRMU2013-70
Volume (vol) vol.113
Number (no) 346
Page pp.pp.-
#Pages 6
Date of Issue