Comparative Analysis of Deep Learning Models for Currency Recognition and Value Detection
Tata Kaushik, Aasritha Sri Vani, N Vithyatharshana, Meena Belwal, Sangita Khare · 2024
Accurate currency recognition and conversion are essential for smooth cross-border transactions in today’s globalized economy. Yet, the wide variety of cash denominations and variances, combined with the complexities of deep learning and image processing, make this work extremely difficult. A comparative comparison of deep learning models is helpful in addressing these issues by identifying the best method for value identification and currency recognition. The objective of this research is to examine how well ResNet50V2, VGG16, and MobileNet perform in terms of precisely identifying and valuing currencies. Utilizing a sample of Thai and Indian banknotes from IEEE data port that include notes with differences in orientation (rotated at 180-degree angles), lighting conditions, and other environmental factors., the study assesses how well each model handles the complexity of money data. The purpose of this study is to give practitioners in the banking and finance industries useful insights.