Optical Character Recognition for Indian Language Scripts Using Support Vector Machines.

Date of Submission

December 2003

Date of Award

Winter 12-12-2004

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science

Department

Computer Vision and Pattern Recognition Unit (CVPR-Kolkata)

Supervisor

Mitra, Mandar (CVPR-Kolkata; ISI)

Abstract (Summary of the Work)

The basic aim of this projevt is to find different methods for the Improvetment of an existing Bangla OCR system. This can be attempted in two ways. One is at the feature selection level; the other is the classification algorithm. We take the latter approach as the current feature is giving reasonably good results. We tried various post processing methods for pairs of charac- ters which are often confused by the current nearest neighbour (NN) scheme. Almost all these methods gave good results cơmpared to the NN method. We have also used Support Vector Machines (SVMS) for postprocessing, as well as for charactet recognition. While the perfor- mance of SVMS is proinising in the postprocessing phase, its performance as a claşsification method is lower than the sinmple NN-hnst appronch. Furthér study is necessary to see the usefulness of SVMS as a classfication scheme.

Comments

ProQuest Collection ID: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:28843444

Control Number

ISI-DISS-2003-118

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/6289

This document is currently not available here.

Share

COinS