Affiliations: [a] Department of I. T. St. Thomas' College of Engg. & Tech., Kol, India | [b] CVPR Unit, I.S.I. Kol, India | [c] Department of I.T., Solt Lake Campus, Jadavpur University, Kol, India
Correspondence:
[*]
Corresponding author: Ranjit Ghoshal, Department of I. T. St. Thomas' College, 4, D. H. Road, Kol-700023, India. E-mail:[email protected]
Abstract: This article proposes a scheme for automatic recognition of Bangla text extracted from outdoor scene images. For extraction, first the headline is obtained, then certain conditions are applied to distinguish between text and non-text. By removing the headline, the Bangla text is partitioned into two zones. Further, an association among the text symbols in these two different zones is observed. For recognition purpose, a decision tree classifier is designed with Multilayer Perceptron (MLP) at leaf nodes. The root node takes into account all possible text symbols. Further nodes highlight distinguishable features and act as a two-class classifiers. Finally, at leaf nodes, a few text symbols remain, that are recognized using MLP classifiers. The association between the two zones makes recognition simpler and efficient. The classifiers are trained using about 7100 samples of 52 classes. Experiments are performed on 250 images (200 scene images and 50 scanned images).
