JAIST Repository >
School of Information Science >
Conference Papers >
Conference Papers >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10119/12347

Title: Iterative identification framework for robust hand-written digit recognition under extremely noisy conditions
Authors: Lee, Hosun
Jeong, Sungmoon
Matsumoto, Tadashi
Chong, Nak Young
Keywords: Improved Viterbi algorithm
Sequential image encoding
Issue Date: 2014-08-18
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Magazine name: 2014 IEEE International Conference on Automation Science and Engineering (CASE)
Start page: 728
End page: 733
DOI: 10.1109/CoASE.2014.6899409
Abstract: A new classification framework for noise invariant hand-written digit recognition is proposed, which is based on Turbo decoding technique and Viterbi algorithm known from communication field. The digit image is modeled twodimensionally correlated Markov Chain Model (MCM) and iteratively exploited on horizontal and vertical direction. In order to increase discriminant of Markov Chain models foreach digit, a novel sequence learning algorithm is proposed to obtain sequence map and Markov chain models which represent the each class by minimizing an entropy information of MCMs within individual digit classes. The effectiveness of the proposed algorithm is verified through experiments that can significantly enhance the character classification accuracy under several noise conditions.
Rights: This is the author's version of the work. Copyright (C) 2014 IEEE. 2014 IEEE International Conference on Automation Science and Engineering (CASE), 2014, 728-733. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
URI: http://hdl.handle.net/10119/12347
Material Type: author
Appears in Collections:b11-1. 会議発表論文・発表資料 (Conference Papers)

Files in This Item:

File Description SizeFormat
20324.pdf338KbAdobe PDFView/Open

All items in DSpace are protected by copyright, with all rights reserved.


Contact : Library Information Section, Japan Advanced Institute of Science and Technology