Physical Currency Detector for Visually Impaired Person

Physical Currency Detector for Visually Impaired Person

Prof. Pallavi Chaudhari1, Prathamesh Mate2, Gaurav Wani3, Pritesh Wadile4, Pranav Patil5

1 Professor, Department of Computer Engineering, Sandip Institute of Technology & Research Centre, Nashik

2,3,4,5 Student, Department of Computer Engineering, Sandip Institute of Technology & Research Centre, Nashik

Abstract—Banknote identification systems are one of the most widely researched fields today, with widespread applications in Automated Teller Machines (ATMs), vending machines, and currency recognition aids for the visually impaired. Using a modular approach, the present work proposes a novel technique for recognizing Indian currency banknotes. The suggested approach extracts special characteristics of notes, such as the center numeral, RBI stamp, color strip, and other marks, and uses algorithms that are designed to recognize each distinct characteristic. The proposed method is testing on a large data set for recognizing Indian banknotes of various denominations and lighting conditions, such as new notes, wrinkled notes, and non-uniform illumination. We propose a convolutional neural network model that considers many images of various denominations. The specimen photographs are captured in a variety of situations and environments. This work will eventually be expanded into a tiny, portable handheld device.

Keywords— neural network, currency, deep learning, visually impaired, CNN