A Comparative Study Of Fake Indian Currency Detection Techniques (KEY IJP************758)

• Pasham Avani,Darbasayanam Sharvani ,Mergu Anudeep ,Vemula Jyothi ,Dr. V. Ravi Kumar

Abstract

The Counterfeiting of Indian currency poses a significant threat to the economy and public trust, necessitating the development of efficient counterfeit detection systems. This study introduces a novel approach leveraging Convolutional Neural Networks, a powerful class of deep learning models renowned for their exceptional image processing capabilities, MobileNet and ResNet. The proposed system initiates by acquiring high-resolution images of banknotes, which undergo preprocessing to extract crucial features such as watermark patterns, security threads, and serial numbers. The architecture is then trained on an extensive dataset comprising genuine and counterfeit currency samples, enabling the model to effectively differentiate between authentic and fake banknotes based on these distinctive features. It begins by acquiring high-resolution images of banknotes, which are then preprocessed to extract crucial features such as watermark patterns, security threads, and serial numbers. These features are key indicators of authenticity and are used by the model to differentiate between genuine and counterfeit banknotes. Experimental results showcase the system's robustness and high accuracy in identifying counterfeit currency, positioning it as a promising tool. The research contributes to ongoing efforts to safeguard the integrity of Indian currency and uphold public trust in financial transactions.Keywords: Currency image dataset, CNN, MobileNet, ResNet