Classification of Sassanin Coin Images Using Statistical Features Fusion

In recent years significant efforts have been made to provide an automated classification system for the classification of ancient coins. The process of classifying ancient coins is faced with numerous problems that algorithms like SIFT are not able to solve all of them. The purpose of this paper is toprovide a method for classifying ancient coins using a combination of statistical features. The statistical features fusion extracted from BDPCA, DWT and DCT transforms by producing suitable feature vectorwith smaller size improves the overall performance of the system. To increase the accuracy rate of theproposed system the signs engraved on the coins that have not been considered in previous work, are addressed. The performed experiments present 88.17% of accuracy rate in the recognition process with applying this method on 7 databases of the Sassanin coins’ images

کلیدواژه ها:

ancient coins classification ، support vector machine ، statistical feature fusion