Smart Electronic Real Time Text Recognition Application

Abstract

Abstract

The area of application of optical character recognition has been expanded in the last few years. There are many techniques which exist and provide high recognition rate, but it consumes time in training the network. Various mobile applications have been designed for optical character recognition (OCR) in Android platform and iOS platform. The limitation of such application is computation ability of mobile processor. This paper proposes a framework of optical character recognition on a mobile device using android application. Most of the existing android application works on existing image in the storage or live camera based text recognition system i.e. offline text recognition application. Existing android applications like camscanner, text fairy, text scanner, smart lens etc. does the optical character recognition on existing image stored in storage of android device. Android Application ‘ABBYY’ does text recognition on live camera. All these applications work on the principal of offline text recognition that has limitation of text recognition rate or accuracy rate. In this work, we represent a neural network based approach to reduce the time and to maintain high accuracy rate. The main idea is to design an android application which works on already trained neural network, which converts real time handwritten text on android device’s touch screen into editable text format. Furthermore sharing the editable text to Arduino Microcontroller through Global System for Mobile and Bluetooth for multipurpose.

Keywords: Offline text recognition, online text recognition, OCR, sliding window, Arduino, GSM, bluetooth, Android studio

Cite this Article

Ramesh Chauhan, Dhaval Pipalia. Smart Electronic Real Time Text Recognition Application. Journal of Electronic Design Technology. 2017; 8(3): 1–7p.