Presentation 2002/5/10
Extraction and Recogniton of Hangul from Scenery Images
Daewoo KIM, Hiroki TAKAHASHI, Masayuki NAKAJIMA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The paper proposes a method which extracts Hangul regions and recognizes Hangul alphabets in images captured by a digital camera. The proposed method extracts candidates of Hangul alphabets from edge images by using the feature of Hangul structure. As a subsequent process, based on the Hangul structure and homochromaticity of an identical text, the Hangul regions are extracted by clustering procedure in color space. Texture regions and noise are excluded and Hangul alphabet regions are extracted from edge images by using simplicity of the Hangul structure. The extracted candidates are classified into character regions and background in color space. The clustering process is performed in circumscribed rectangles of extracted edges in the previous procedure. Moreover, the proposed method recognizes the extracted Hangul alphabets which are ideally written. It analyzes endpoints and intersected points of the extracted alphabet skeletons. The proposed method extracts 85.97% of Hangul text from 100 images captured on arbitrary conditions. Furthermore, 71.96% of alphabets are recognized for ideal Hangul of which Hangul alphabets are not connected each other in a character.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Hangul text / Text extraction / Character recognition / Digital Camera
Paper # PRMU2002-18
Date of Issue

Conference Information
Committee PRMU
Conference Date 2002/5/10(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) Extraction and Recogniton of Hangul from Scenery Images
Sub Title (in English)
Keyword(1) Hangul text
Keyword(2) Text extraction
Keyword(3) Character recognition
Keyword(4) Digital Camera
1st Author's Name Daewoo KIM
1st Author's Affiliation Graduate School of Information Science & Engineering, Tokyo Institute of Technology()
2nd Author's Name Hiroki TAKAHASHI
2nd Author's Affiliation Graduate School of Information Science & Engineering, Tokyo Institute of Technology
3rd Author's Name Masayuki NAKAJIMA
3rd Author's Affiliation Graduate School of Information Science & Engineering, Tokyo Institute of Technology
Date 2002/5/10
Paper # PRMU2002-18
Volume (vol) vol.102
Number (no) 55
Page pp.pp.-
#Pages 6
Date of Issue