Fake Currency Detection Using Convolutional Neural Networks

D. Jahnavi; P.M.K. Prasad; Guntu Nooka Raju; BPV Dileep; K. Lohitha

Abstract

In today’s world, due to increasing technology like scanning, color printing, and duplicating, the identification of bogus notes by the human eye is almost getting impossible. Knowingly or unknowingly, due to the usage of bogus currency notes, the Indian economy is also being impacted badly. Hence, the identification of bogus currency notes is really important. This paper deals with regard to identifying whether the given sample of the currency note is real or bogus. Previously, there are many methods for the identification of bogus notes eliminating the need for manual feature extraction. Also, the results are not very accurate and hence deep neural networks can be used for the recognition of bogus currency notes. In this paper, the Convolution Neural Network technique is employed for the detection of bogus currency bills. Convolution Neural Network is an architecture that directly learns from the input data and classifies the note, if the note is found to be a bogus note, then immediately a message will be sent indicating the specified location where the bogus notes are being circulated. Hence, the problem of bogus note circulation can be reduced and also help the society not being deceived by fraudsters. The results shows that the training accuracy and validation accuracy of the proposed method are 99.6% and 99.7% respectively

Abstract

Keywords