Published Paper Details:

Currency counterfeiting has become one of the threats to the economy of a country. This is more severe in cash based economies while systems relying on more sophisticated forms of currency are less prone to loses due to counterfeiting. For any of the economies to be stable, there universality of acceptance of one standard currency. This in turn ensures absence of the phenomenon of currency counterfeiting. Such techniques how ever are expensive as well as labor intensive. The primary focus of this paper then is to analyze deep learning algorithms, specifically GAN Ensemble Models, in the counterfeiting of Indian currency. However, the methodology proposed in this paper has put aside economics, and focuses instead on physics and computer vision techniques such as Contrast Limited Adaptive Histogram Equalization (CLAHE). It focuses on enhancing image quality in an efficient manner. Experimental results reveal that the accuracy achieved by GAN networks enhanced by CLAHE surpassed the results achievable from regular CNN models with a margin of more than percent which means that such methods are capable of being implemented in practice as viable and cost effective solutions to the emerging problem of counterfeiting.