A Robust Invariant Image-Based Paper-Currency Recognition Based on F-kNN

Abstract

The innovation of currency recognition intends to look, distinguish and remove the noticeable just as imperceptible subtleties on paper-money for effective classification of currency. Many a times, currency notes are hazy or harmed; a considerable lot of them have complex structures as well. This makes the assignment of currency recognition troublesome. Currency recognition is applied in order to diminish the human influence, put resources into this procedure. So it is essential to choose the correct highlights and legitimate calculation for this reason. This work presents a framework for automated currency notes recognition utilizing supervised image processing strategies. This work is critical considering the mentioned dimensions, namely, a) They got worn-out ahead of their schedule in comparison to coins; b) The possibility of joining wear-out currency is more noteworthy than that of coin currency; c) Coin currency is restricted to lesser population. We have to actualize a calculation which should be straightforward, less mind-boggling and profoundly effective. Recognition of Paper-Currency is significant in the zone of pattern recognition. Image processing is used to acquire the final outcome, with accuracy of 51.0%, 56.8%, 65.6% for the Decision Tree, SVM, Fine-KNN classifiers respectively.