Fake Currency Identification Using Artificial Intelligence and Federated Learning
Syed Zahiruddin, Vamsi Krishna Kadiri, Valli Bhasha Achukatla, Pavan Kumar Kattela, Ahmed A. Elngar · 2024
Avoiding the generation and use of fake currency is a problem in the entire world. It has become a significant problem that has adversely affected almost every nation. Identification of paper currency is the main intent of this chapter. The data set is collected, and the chapter currency of various denominations is taken, including both original and fake notes. For identifying and classifying the currency, artificial intelligence and federated learning are applied; first, a network is built by outlining the architecture using a convolutional neural network (CNN) and stochastic gradient descent with momentum (SGDM) optimization. Then the network is trained with data sets. For both identification and classification, the CNN network is used, and also determining whether the currency is original or counterfeit is accomplished using the MATLAB image processing toolbox. The trained network average accuracy determined is 95.625%. In terms of identification, classification, and accuracy detection, the results of the suggested technique are found acceptable.