Recognition of characters in Indian MODI script

Modi was very useful script in the kingdoms of medieval Maharashtra. In the reign of the great Maratha-Chhatrapati Shivaji and also in the reign of Peshwas, this script was widely incorporated in ruling the state. This script is very similar to the shorthand. At that time, it was used in Maharashtra to prepare the documents such as Property matters, Donation of Land (Dan-Patra), Land Revenue, Military orders (Farman), Justice, Religious matters, Jahagir vatan, Judiciary, etc. In this research, Analysis of the MODI script is studied. Identification and Recognition of the handwritten MODI characters are performed. The database of handwritten samples is prepared by using ANESP program. The original manuscript of MODI script is collected and it is preprocessed. Moment Invariant and Affine Moment Invariant are the two techniques which are used for extracting the features from the handwritten segregated samples. Identification and recognition is done using machine learning technique. Among the variants of machine learning techniques, support vector machine is used as a classifier. Linear kernel function is applied in this support vector machine while performing classification. This system provides good recognition rate for the handwritten samples of MODI script. This research will be helpful to the researchers to enlighten the uncovered part of history.