An Image Processing-Driven System for Fake Currency Detection

Abstract

Fake currency is a critical factor affecting economies worldwide, including India. In this paper, a novel and structurally efficient approach for detecting and identifying duplication in currency notes is presented using Discrete Wavelet Transform (DWT). The system employs image processing algorithms to extract essential features such as security thread, intaglio printing (RBI logo), and identification mark, which serve as security measures for Indian currency. To identify fake portions in the currency notes and make informed decisions about their authenticity, the matching scores from all fake detection modules are fused together. A crucial aspect of the work lies in comparing the extracted features from various currency notes, enabling us to differentiate between fake and genuine notes effectively. To assess the performance, mean square error is employed as a metric for comparison between two images. A database is build containing authentic Indian notes of different denominations, extract their features, convert them into binary equivalents, and then calculate their mean square error. The proposed Fake Note Detection System takes a test currency note image, performs preprocessing operations to eliminate noise and negative artifacts, and then proceeds with the detection process. The system offers a promising solution to combat counterfeit currency issues and safeguard the integrity of the Indian economy.