A Complete Tamil Optical Character Recognition System
DAS '02 Proceedings of the 5th International Workshop on Document Analysis Systems V
On Developing High Accuracy OCR Systems for Telugu and Other Indian Scripts
LEC '02 Proceedings of the Language Engineering Conference (LEC'02)
An optical character recognition system for printed Telugu text
Pattern Analysis & Applications
| Hi-index | 0.00 |
Efficient algorithms in the sense of reliability and speed for pattern classification and character recognition effectively continue to be a challenge today in newer sub-spaces. There is a large sector of challenging offline or near- real time applications in optical character recognition. Images are represented as matrices in some form or other and characterization of these image matrices for quick classification is a rather demanding task. In this paper we propose two methods for characterization one based on computation of marginal frequencies and the other uses dominant singular values as distinct characteristic features. This feature selection method reduces the computational complexity without compromising efficiency. Results and analysis including the comparison of the efficiency of the methods have been presented.