Large Scale Font Independent Urdu Text Recognition System

Abstract

OCR algorithms have received a significant improvement in performance recently, mainly due to the increase in capabilities of artificial intelligence algorithms. However this advancement is not evenly distributed over all languages. Urdu is among the languages which did not receive much attention, specially in the font independent perspective. There exists no automated system that can reliably recognize printed Urdu text in images and videos across different fonts. To help bridge this gap, we have developed Qaida, a large scale data set with 256 fonts and complete Urdu lexicon. We have also developed Convolutional Neural Network (CNN) based classification models which can recognise Urdu ligatures with 84.2% accuracy. Moreover we demonstrate that our recognition network can not only recognise the text in the fonts it is trained on, but can also reliably recognise text in unseen (new) fonts. To this end, this paper makes following contributions: (i) we introduce a large scale, multiple (more than 200) fonts based data set for printed Urdu text recognition 1 ; (ii) we have designed, trained and evaluated a CNN based model for Urdu text recognition; (iii) we experiment with incremental learning methods to produce state-of-the-art results for Urdu text recognition. All the experiment choices were thoroughly validated via detailed empirical analysis. We believe that our this study can serve as the basis for further improvement in the performance of Urdu OCR systems.