Recognition of Devanagari Scene Text Using Autoencoder CNN


  • Sankirti Sandeep Shiravale Marathwada Mitra Mandal's College of Engineering, Pune
  • Jayadevan R Army Institute of Technology, Pune
  • Sanjeev S Sannakki Gogte Institute of Technology, Belagavi


Scene text recognition is a well-rooted research domain covering a diverse application area. Recognition of scene text is challenging due to the complex nature of scene images. Various structural characteristics of the script also influence the recognition process. Text and background segmentation is a mandatory step in the scene text recognition process. A text recognition system produces the most accurate results if the structural and contextual information is preserved by the segmentation technique.  Therefore, an attempt is made here to develop a robust foreground/background segmentation(separation) technique that produces the highest recognition results. A ground-truth dataset containing Devanagari scene text images is prepared for the experimentation. An encoder-decoder convolutional neural network model is used for text/background segmentation. The model is trained with Devanagari scene text images for pixel-wise classification of text and background.  The segmented text is then recognized using an existing OCR engine (Tesseract). The word and character level recognition rates are computed and compared with other existing segmentation techniques to establish the effectiveness of the proposed technique.


Character and Text recognition, scene text recognition, Devanagari script, OCR, segmentation technique, encoder-decoder CNN

Author Biographies

Sankirti Sandeep Shiravale, Marathwada Mitra Mandal's College of Engineering, Pune

Sankirti S. Shiravale has completed B.E degree from Shivaji University, Kolhapur in 2003, M. E. degree from University of Pune, Pune in 2012. Currently, she is with Department of Computer Engineering, Marathwada Mitra Mandal’s College of Engineering, Pune as an Assistant Professor and pursuing her research from VTU, Belagavi. Her research area is Image Processing and Pattern Recognition.

Jayadevan R, Army Institute of Technology, Pune

Jayadevan R received the B. Tech degree from Cochin University of Science and Technology, Kochi in 2002, the M.E. degree from University of Pune, Pune in 2006, both in Computer Science and Engineering, and the Ph.D. degree in Computer Engineering from North Maharashtra University, Jalgaon, in 2013. He is currently an Associate Professor with the Department of Computer Engineering, Army Institute of Technology, Pune, India. His research interests include Image Processing and Pattern   Recognition.

Sanjeev S Sannakki, Gogte Institute of Technology, Belagavi

Sanjeev S Sannakki, has completed his Ph.D. degree in Image processing & Data Mining from VTU Belagavi. His career spans over a period of two decades is in the field of teaching, research and other diversified in-depth experience in academics. He is currently working as a Professor in the Department of CSE, Gogte Institute of Technology, Belgaum. Currently he is shouldering the responsibility of Head of the Research centre. He has published several papers in reputed national/international conferences and journals. He is also guiding the research scholars & UG/PG students of VTU.                 




Download data is not yet available.