A Two-stage Character Segmentation Technique for Printed Kannada Text

Due to the presence of bottom extension characters (Vatthus), adjacent characters in a Kannada word sometimes overlap in the vertical projection profile. In such cases, the usual method of character segmentation using projection profile is not efficient. In this paper we present a segmentation method in which the Vatthus of a Kannada word are segmented first by using connected component algorithm, and then the remaining characters of the word are segmented by vertical projection profile method. The segmentation method was tested on a large data set of 25000 words and a segmentation rate of 98.2% is archived.

Keywords:

Kannada text, Projection profile, Character segmentation, Connected   components.