Writer identification for offline Tamil handwriting based on gray-level co-occurrence matrices

Abstract

Writer identification is a popular research field in many languages. Since alphabets of different language have different pattern, the methods are dependent on the language. Handwriting of a person has some features which are unique to every person and can be used for identification. Most of the research activities for writer identification are based on English documents and the research activities for Tamil handwriting are thin and databases are not available. In this paper, a method is proposed to identify a writer from the scanned images of Tamil handwritten text. Our approach is based on texture analysis where each writer's handwriting is treated as a different texture. The method is text independent and based on the features extracted from gray level co-occurrence matrix of the scanned image. Handwriting samples from 70 writers were used and the documents were scanned with 150 dpi.