Result Paper on CurrencyShield: Deep Learning for Financial Security

Abstract

Counterfeit money is a big problem for the economy, causing people to lose trust and affecting financial stability. This paper presents a Fake Currency Detection System designed for Indian banknotes, using advanced deep learning and image processing to accurately tell the difference between real and fake notes. The system uses a Convolutional Neural Network (CNN) to check important features on currency notes, like the image of Mahatma Gandhi, serial numbers, and security strips, which help to identify if a note is genuine. By focusing on these security features, the system can detect small changes or missing details that are common in counterfeit notes. The system is available as a web-based app, built with Python and Flask, where users can easily upload pictures of currency notes for checking. After submitting an image, the CNN model will quickly decide if the note is real or fake. To make it easier for everyone to use, the system offers voice feedback in English, Hindi, and Marathi, which is especially helpful for people with visual impairments. This feature ensures that people from different language backgrounds in India can use the system easily.