Isolated Telugu Palm leaf character recognition using Radon Transform — A novel approach

This paper deals with the Palm leaf character recognition (PLCR) using Radon Transform applied to Telugu Characters. A large collection of these Palm leaf characters are available in the classical Indian languages like Sanskrit, Tamil, Pali etc as well as in more modern languages like Telugu. Manuscripts on Palm leaves in India are the most unique collection for Centuries pertaining to wisdom and knowledge containing religious texts and treaties on a host of subjects such as art, medicine, astronomy, astrology, mathematics, law and music in various traditional and modern languages. The palm leaves are natural organic products and are therefore very susceptible to deterioration due to climatic factors (relative humidity, temperature), light and insects. Hence, preservation and digitization of these palm leaves/manuscripts is important. These characters on the palm leaf have the additional properties like depth (which is proportional to the pen pressure applied by the scriber), an added feature which can be gainfully exploited in PLCR. This paper explores how these 3D features can be extracted and how they can be gainfully used in the recognition and classification process using Radon Transform and Nearest Neighborhood Classifier (NNC). The best percentage of accuracy obtained in the proposed method is 93%.