Currency Recognition System for Visually impaired using a Novel CNN-LSTM based Hybrid Approach

Abstract

One of the main problems faced by visually impaired people is dealing with monetary transactions, due to their inability to recognize the value of currency notes because of the identical dimensions and it’s feel . We introduce in this paper, a camera phone-based money recognition system that executes real-time processing for every detected frame as the camera advances towards the note using CNN-LSTM algorithm and then generates a corresponding audio message telling about the value of the currency note. To make the system reliable in real-life scenarios, we have created a Currency recognition system (CRS) built on Convolutional neural network (CNN) and Long short-term memory (LSTM). The hybrid CNN-LSTM algorithm obtains results as fast and robust as possible.