Individuality of Isolated Bangla Numerals

Authors

  • Chayan Halder 1Department of Computer Science, West Bengal State University
  • Kaushik Roy

Keywords:

Individuality of Handwriting, Writer Identification, Bangla Handwriting, WEKA, MLP, LIBLINEAR

Abstract

Writer Identification and Verification is a study in the field of Computer Vision and Pattern Recognition. The identification and verification of writer through the analysis of handwriting has significant role for many critical judicial decisions. The forensic experts often encounters this type of problem to identify or verify writer from given documents or part of documents. In todays high-tech world an automated system with the ability to identify and verify writers can play vital role in the judicial system. In current times, to the best of our knowledge there is no such automated system of Writer Identification and Verification on Indic script like Bangla, Oriya, Telugu etc. Analysis of individuality of handwritten isolated Bangla numerals are presented here. It has a great prospect not only in Writer Identification and Writer Verification but also in Graphological Analysis and also in various fields of Forensic Science based
on handwritten documents etc. As there is no such standard Bangla writer database, we have collected data samples consisting of total 4500 numerals from 90 writers with 5 sets from each writer. After collecting and extracting characters from filled in forms, 64 and 400 dimensional feature vectors are computed on numerals based on directional chain code and gradient of the images. In our experiment we have used LIBLINEAR and MLP classifier of WEKA environment. We have computed and analyzed the Individuality of each numeral and observed that the numeral PANCH (5) is the most individual than other numerals except when MLP classifier is used for classification on 400 dimensional feature set where numeral DUI (2) is the most individual. It has also been observed that numeral SUNNO (0) has the least individuality. We have also done the writer identification with all the numerals and using 64 dimensional feature we have obtained 97.07% accuracy for LIBLINEAR classifier and 94.62% accuracy for MLP classifier with all writers. For 400
dimensional feature with LIBLINEAR classifier 96.5% writer identification accuracy has been achieved.

Downloads

Download data is not yet available.

Downloads

Published

2013-01-01

How to Cite

Chayan Halder, & Kaushik Roy. (2013). Individuality of Isolated Bangla Numerals. Journal of Network and Innovative Computing, 1, 10. Retrieved from https://cspub-jnic.org/index.php/jnic/article/view/14

Issue

Vol. 1 (2013)

Section

Original Article

License

You are free to:

  1. Share — copy and redistribute the material in any medium or format for any purpose, even commercially.
  2. Adapt — remix, transform, and build upon the material for any purpose, even commercially.
  3. The licensor cannot revoke these freedoms as long as you follow the license terms.

Under the following terms:

  1. Attribution — You must give appropriate credit , provide a link to the license, and indicate if changes were made . You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
  2. No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.

Notices:

You do not have to comply with the license for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation .

No warranties are given. The license may not give you all of the permissions necessary for your intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you use the material.