Real and Fake Currency Detection using ANN

Currency is the main form of exchange in India and is essential to the country’s financial system, social
progress, and economic growth. In today’s highly technologically advanced culture, counterfeit money is a
major problem since paper money is easy to move and safe to have, but its face value is significantly higher
than its actual value.
Acknowledging the significance of preserving economic advancement, the Indian government has implemented
policies like the demonetization of the Rs. 1000 and Rs. 500 notes. Though it becomes a possible target for
counterfeiters, the introduction of the Rs. 200 note and the revised design for the Rs. 500, Rs. 100, Rs. 50,
Rs. 20, and Rs. 10 notes provide new problems. The main problem with the hardware-based methods now
in use for detecting counterfeit notes is that they are challenging for average people to use, even with their
competence.
This essay looks at the characteristics that set the new legal tender from the Reserve Bank of India apart
and uses methods to identify and confirm that it is real. a hybrid method that combines an ANN with a ResNet
model, some architecture, and ANN adjusted parameters to accurately identify counterfeit money. This method
employs residual networks (ResNet 50) and artificial neural networks (ANN) to recognize counterfeit currency
based on its width, color, and serial number. To distinguish authentic notes from fakes, the system goes
through a number of processes, such as pre-processing, segmenting, comparing, and extracting attributes from
pictures. Finding counterfeit money is still a difficult and complex process. General Terms: Image Processing,
Feature Extraction, Detection.