Reconstructing strokes and writing sequences from chinese character images

Research output: Chapters, Conference Papers, Creative and Literary Works (RGC: 12, 32, 41, 45)32_Refereed conference paper (with ISBN/ISSN)peer-review

3 Scopus Citations
View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Original languageEnglish
Title of host publicationProceedings of the Sixth International Conference on Machine Learning and Cybernetics, ICMLC 2007
Pages160-165
Volume1
Publication statusPublished - 2007

Publication series

Name
Volume1

Conference

Title6th International Conference on Machine Learning and Cybernetics, ICMLC 2007
PlaceChina
CityHong Kong
Period19 - 22 August 2007

Abstract

The Chinese characters evolved from pictograms and they are composed of strokes. A standard stroke sequence for each character is available in the dictionary. People introduced heuristic rules to specify the stroke order for easy memorization but it is very ambiguous to reconstruct the dictionary sequence according to the heuristic rules. In this paper, we combine the stroke extraction and stroke sequence reconstruction algorithms to reconstruct the strokes and their sequence from a Chinese character image. A well-known public Chinese character database (the HITPU database) is used as our input data. Performance evaluation shows the robustness of our proposed method and user evaluation shows that our proposed system helps users to create online Chinese character templates quickly and conveniently. ©2007 IEEE.

Research Area(s)

  • Chinese character, HITPU database, Pattern classification, Stroke extraction, Stroke sequence estimation

Citation Format(s)

Reconstructing strokes and writing sequences from chinese character images. / Tang, Kai-Tai; Leung, Howard.

Proceedings of the Sixth International Conference on Machine Learning and Cybernetics, ICMLC 2007. Vol. 1 2007. p. 160-165 4370133.

Research output: Chapters, Conference Papers, Creative and Literary Works (RGC: 12, 32, 41, 45)32_Refereed conference paper (with ISBN/ISSN)peer-review