Currency Detection and Recognition System Based on Deep Learning

Currency Detection and Recognition System Based on Deep Learning

Ms. P. Divya Jenifar¹,Ms. V. Harinitha², Ms. M.D. Harshini³,Ms. A. James Christiya⁴

¹Assistant professor ,Department of Biomedical Engineering,Sri Shakthi Institute of Engineering and Technology,Tamilnadu,India.

^2,3,4Final  year  students,Department of Biomedical Engineering,Sri Shakthi Institute of Engineering and Technology,Tamilnadu,India.

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract -This paper suggests a method for recognizing currency through image analysis. The method relies on the three characteristics of color, height, and texture, which utilized for distinction. The method put forth in this document can be used to find the currencies of numerous nations. For the purposes of execution, only paper currencies from India are taken into account. This system, which employs the CNN (Convolution Neural Network), makes it simpler to check currency anywhere, at any moment. Every denomination of India was used in our testing, and the system was able to identify with 95% precision. To improve the technique's accuracy, a classification model is created using all of the previously stated factors. Features of paper money are crucial to this identification.

Key Words: Currency recognition, color, CNN, Open CV, ResNet.