Authors:
Tun Zarni Aung, Sa Kaung Min Htet, Hlaing Htake Khaung Tin
Addresses:
Faculty of Information Science, University of Information Technology, Yangon, Myanmar.
Abstract:
Financial institutions, governments, and businesses around the world face serious challenges because of the growing sophistication of counterfeit currency. Visual inspection and UV light verification are two traditional detection techniques that often miss sophisticated forgeries and require extensive human intervention. Machine learning (ML) and artificial intelligence (AI) techniques have become popular in recent years as viable means of automated, precise, and scalable counterfeit detection. This review paper offers a thorough examination of AI- and ML-based methods for detecting counterfeit currency, including hybrid approaches, image processing, feature extraction, and classification algorithms. Researchers assess how well convolutional neural networks, deep learning architectures, and supervised and unsupervised learning techniques detect counterfeit currency. The study examines the drawbacks of existing technologies, including reliance on datasets, variability in currency design, and difficulties with real-time implementation. This review outlines the advantages, disadvantages, and prospects for future integration of intelligent systems in the fight against currency counterfeiting by contrasting conventional, digital, and AI-driven approaches. According to the results, AI- and ML-based detection systems perform notably better in terms of accuracy and efficiency than traditional techniques, laying the groundwork for safer financial operations. To improve detection capabilities in dynamic financial environments, this review also highlights the need for cross-currency applications, large-scale datasets, and adaptive models.
Keywords: Counterfeit Currency Detection; Image Processing; Machine Learning (ML); Convolutional Neural Networks (CNNs); Financial Security; Feature Extraction; Classification Algorithms; Supervised Learning.
Received: 14/09/2024, Revised: 30/10/2024, Accepted: 04/01/2025, Published: 05/09/2025
DOI: 10.64091/ATICL.2025.000230
AVE Trends in Intelligent Computer Letters, 2025 Vol. 1 No. 3 , Pages: 161-169