Currency Recognition App for Blind People

Visually impaired people face a lot of issues in recognizing currency notes. Although the use of visa, master cards, UPI and other electronic payments is expanding rapidly, paper cash is still widely used for casual exchanges due to its convenience. However, people with visual impairments may find it difficult to identify different currencies. The Currency Recognition App (CRA) can be used to help visually impaired people who face difficulty in dealing with currency. In this research paper, a currency detection App is built based on the proposed CNN (Convolutional Neural Network) algorithm, that facilitates the visually impaired people to recognize the different denominations of Indian Currencies with the basic concept being the usage of Deep Learning Algorithm. A Convolutional Neural Network model has been built in which a high volume of images of different denominations have been considered. There are around 11 different currencies in India, and each banknote differs from the others in a number of ways, including color and size. The specimen images have been captured in different positions and in physical conditions. The proposed model is compared with the present system and the test conducted shows that the proposed CNN is better with an accuracy of 89% which is more when compared to the existing system ANN which has accuracy of less than 71%. The solution to the same is provided by developing an app where once the image is captured, processes the image, says the result what kind of currency it is so that they can hear. It strives to deliver quick findings that are correct for the provided picture. The proposed plan aims to simplify life for someone who is sight impaired.