An Efficient Period Prediction System for Tamil Epigraphical Scripts Using Transductive Support Vector Machine

Abstract: Tamil is one of the ancient languages in the world with rich in literature. The writers used various materials like stone, metal, pottery, wood, palm leaves, cloth, conch shell, mural paintings and copper plates to encrypt their writing. The information gathered from these inscriptions gives us knowledge about the astronomy, history, culture, religious, economic tax, administrative and educational conditions. These epigraphical inscriptions plays an important role in knowing the civilized past and classification of characters belonging to various periods. Therefore a system is proposed to read the ancient Tamil characters belonging to various periods by testing a small amount of characters referred to as examined characters in Tamil language. These examined characters are taken from the script automatically and coordinate with the characters belonging to different periods using machine intelligence. Hence the proposed system consists of various modules like image acquisition, binarization, preprocessing, feature extraction, segmentation, and at last classification and prediction of period using Transductive Support Vector Machine (TSVM).The experimental results shows higher accuracy when compared with Support Vector Machine (SVM)

Keywords: Epigraphical Scripts, Tamil Language, TSVM, Feature Extraction