By Topic

Keyword Extraction from Handwritten Chinese Document Image Based on Matching and Voting of the Local Topological Structure

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Rui Chen ; Coll. of Comput. Sci., ChongQing Univ., Chongqing, China ; Bin Fang

In this paper, we present a keyword extraction methodology from handwritten Chinese document image based on matching and voting of the local topological structure. In the process, a handwritten keyword image is used as template, from which the local topological structure features of each character pixel are extracted, including crossing number, direction angle, gravity direction and gravity distance. A reference table is established based on these features. In the matching and voting phase, the features of each character pixel of handwritten document image are compared with the ones in the reference table. A vote is made in the document image when the crossing number and direction angle match each other and the voting location is determined using gravity direction and gravity distance. The handwritten keywords with local shape variation, slight rotation and zooming can be extracted from handwritten Chinese document image effectively. The same keywords from a same document can be used as training samples, which are used to model writers based on the signature verification methods to accomplish the goal of writer identification.

Published in:

Pattern Recognition (CCPR), 2010 Chinese Conference on

Date of Conference:

21-23 Oct. 2010