Presentation 2014-06-20
An experiment of Text/Non-text Classification using Conditional Random Fields for Japanese Online Handwritten Japanese Ink Documents
Soichiro Inatani, Truyen Van Phan, Masaki Nakagawa,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose the method based on Conditional Random fields (CRF) for separating text and non-text ink strokes in online handwritten Japanese documents. Text/non-text classification requires for the free handwriting, and contributes to the improvement of text and figure recognition accuracies. Several works using context information reported superior results for text/non-text classification to those without using context information. The purpose of this work is to improve the accuracy of text/non-text classification. To attain our goal, we propose the classification method based on CRF utilizing spatial and temporal context of ink documents effectively. In this experiment, we compare the two methods based on CRF and Markov Random fields (MRF) for classifying ink strokes into text and non-text, and evaluate the performances of methods.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Text/Non-text classification / ink stroke classification / Markov Random fields / Conditional Random fields / CRF
Paper # PRMU2014-34
Date of Issue

Conference Information
Committee PRMU
Conference Date 2014/6/12(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) An experiment of Text/Non-text Classification using Conditional Random Fields for Japanese Online Handwritten Japanese Ink Documents
Sub Title (in English)
Keyword(1) Text/Non-text classification
Keyword(2) ink stroke classification
Keyword(3) Markov Random fields
Keyword(4) Conditional Random fields
Keyword(5) CRF
1st Author's Name Soichiro Inatani
1st Author's Affiliation Department of Information Engineering, Tokyo University of Agriculture and Technology()
2nd Author's Name Truyen Van Phan
2nd Author's Affiliation Department of Information Engineering, Tokyo University of Agriculture and Technology
3rd Author's Name Masaki Nakagawa
3rd Author's Affiliation Department of Information Engineering, Tokyo University of Agriculture and Technology
Date 2014-06-20
Paper # PRMU2014-34
Volume (vol) vol.114
Number (no) 90
Page pp.pp.-
#Pages 6
Date of Issue