Devanagari script based Optical Character Recognition and Text to Speech Generation 

Abstract

Optical character Recognition (OCR) is a conversion of scanned or printed text
images, handwritten text into editable text for further processing. This technology
allows machine to recognize the text automatically. OCR is a field of research in
pattern recognition, artificial intelligence and computer vision.
World is contracting with the growth of mobile phone technology. As the number of
users is increasing day by day, facilities are also increasing. Starting with simple
regular handsets which were used just for making phone calls, mobiles have changed
our lives and have become part of it. Now they are not used just for making calls but
they have innumerable uses and can be used as a Camera , Music player, Tablet PC,
T.V. , Web browser etc . And with the new technologies, new software and operating
systems are required. By observing the growth of Android user, we can say that
Android operating system has met the expectation of mobile user. Android platform
has been good platform for the developers due to its large number of users.
This project aims to combine'these two rapidly emerging fields of technology, and
take the users’ mobile experience to the next level. This android application project
implements OCR to extract the recharge card pin number of two major telecom
companies of Nepal NTC and NCELL, to recharge the balance. This application uses
tesseract engine to extract recharge card pin number. The android application uses the
tess two library.
The project result shows, the accuracy of the detected pin number is affected by the
lighting condition during capturing of image. The highest accuracy result was found
during the sunlight which was approximately 72% and in normal light condition the
accuracy was found to be approximately 54%.