Indian Currency Classification and Fake Note Identification using Feature Ensemble Approach

Abstract

Visually challenged people face a lot of difficulty in currency recognition and counterfeit notes identification. The change of currency notes during different time frames also adds more difficulty in recognizing features in currency notes manually by the visually impaired. Further, the number of counterfeit notes in circulation in India is increasing every year. Computer-aided currency recognition and fake note detection can help people in the identification of currency denomination without any dependency on manual intervention. In this work, we have proposed a currency recognition algorithm that can classify an Indian currency and help in identifying the fake notes without manual intervention and thereby preventing their circulation in the country. In this work, Chan Vese Segmentation is used to segment the security features of a note and then an ensemble of classifiers is used for classification and fake note identification. Experimental results show promising results even on the small dataset. The method can be extended to currency recognition for different countries.