Machine Learning Paradigms for Counterfeit Currency Detection: A Comprehensive Comparative Study
Tejashree Narayan Agasti, Abhiram Gudimella, M. Shanmugasundaram · 2024
This academic paper presents a unique method that uses machine learning and image processing techniques to identify counterfeit Indian currency notes. In the context of the digital age, conventional methods have demonstrated their limitations. The proposed methodology integrates various machine learning algorithms, such as Support Vector Machine (SVM), Random Forest, Logistic Regression, Decision Tree, and Convolutional Neural Network (CNN), employing a comprehensive dataset for banknote authentication. Notably, the Random Forest algorithm yields a remarkable accuracy of $97.7 \%$ for counterfeit currency detection, while Decision Tree exhibits the best performance in terms of wall clock time, thereby underscoring its pivotal role in the realm of fraud prevention.