Implementing Kannada Optical Character Recognition on the Android Operating System for Kannada Sign Boards

Abstract: This paper focuses on an application that performs Kannada OCR (Optical Character Recognition) in hand held devices. The objective is to make use of the visual capabilities of the built in camera of Android devices to extract text from Kannada sign boards, newspapers etc., and use the same platform to implement the OCR technology with the help of kohonen’s algorithm. Another aim of this application is to find the meaning of the word detected on the Internet. With the help of an initial, sample data, the Kohonen network is trained. The image taken is further processed by initially thinning the characters in the image using the Hilditch algorithm. The image is processed, eliminating small distortions present. The image is then converted to into a greyscale image, which is segmented and the result is displayed along with a vocal output. These characters may be edited by the user and its meaning is then displayed in English.

Keywords: Optical Character Recognition, JAVA, JVM, Android, Image Processing