Indian Currency Classification Using Deep Learning Techniques

Abstract

Progression and evolution of technology has superseeded mechanical human workload in almost every domain with the operation of machines. The currency paper recognition is applicable in various domains of automatic selling goods systems and in banking systems. In the modern transition world for the automatic current recurring systems, the precise identification of paper currency notes is indeed an essential need. Machines often find it difficult in identifying and recognising the currencies in the market when the currency notes have turned bleary and damaged. It is hard for visually disabled people without any technological support or assistance to predict and analyze genuine currency notes. The accuracy of currency notes analysis identification have been refined and boosted throughout with the assistance of these models. Our research methodologies are in line and meeting the desired expectations. This paper presents an Indian Paper Currency Prediction Analysis, proposes an optimized model to recognise the currencies effectively. The Deep Learning approach of CNN model technique has improved the effective analysis of currency recognition with improved accuracy, high speed and efficiency along with complete automatic readily procedure with no human intervention and minimal complexity. This paper represents a strategy which is parted into two divisions, Keras trained a DL Model as well as hosted a Flask based web app on Heroku.Our proposed algorithm design and experimental based results are useful for majorly visually impaired people for differentiating all sorts of available denominations.