Optical character recognition of Gurmukhi script using multiple classifiers

Authors:
Gurpreet Singh Lehal
Affiliations:
Punjabi University, Patiala, India
Venue:
Proceedings of the International Workshop on Multilingual OCR
Year:
2009

Citing 8
Cited 1

Character image enhancement by selective region-growing

Pattern Recognition Letters
On Combining Classifiers

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Theoretical Study on Six Classifier Fusion Strategies

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Majority Voting Scheme for Multiresolution Recognition of Handprinted Numerals

ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 1
Combining model-based and discriminative classifiers: application to handwritten character recognition

ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 1
Classifier combination for improved lexical disambiguation

COLING '98 Proceedings of the 17th international conference on Computational linguistics - Volume 1
Automatic accurate broken character restoration for patrimonial documents

International Journal on Document Analysis and Recognition
Application of majority voting to pattern recognition: an analysis of its behavior and performance

IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans

A bilingual Gurmukhi-English OCR based on multiple script identifiers and language models

Proceedings of the 4th International Workshop on Multilingual OCR

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we present a robust and font independent Gurmukhi OCR system, which performs reasonably well on old documents as well. The OCR is based on four classifiers operating in serial and parallel mode. For combining the results of the classifiers operating in parallel mode, a corpus based weighted voting method is used. Combining multiple classifiers in such a way, that their individual weaknesses are compensated while their individual strengths are preserved, results in significantly better performance than what can be achieved with a single classifier. The problem of broken characters, which frequently appear in old documents, has also been tackled using a structural feature based algorithm.