CNN based framework for identifying the Indian currency denomination for physically challenged people

Abstract

One of the premier issues confronting the visual hindered individual is money, acknowledgment especially for cash. Be that as it may, the outwardly weakened individual may not think about the estimation of cash, and they endure part in cash trade related issues in their normal life. To address this issue, we have built up a framework for acknowledgment of money, notes, which might be the helpful device for an outwardly debilitated individual. Investigation and trials were done on the money informational collection, which encouraged CNN dependent on the key highlights, for example, watermarks, pictures printed on cash, esteemed composed as words and numbers and the total cash. This paper deals with the utilization of Convolutional Neural Networks (CNNs) for solving this society issues and investigations about the exhibition and evaluation of different CNN models. Here, Alexnet, Googlenet, and Vgg16 models have been considered for assessment. All the models were adjusted as far as preparing and testing the individuals of data sets. Among these three models, Alexnet accomplished better in preparing fulfillment, Vgg16 model indicated the better execution and accomplished 100%, Google net arrives at 88% as far as productivity.