Ensemble Deep Learning for Brazil Currency Coin Prediction. Issue 1 (February 2021)

• Record Type: Journal Article
• Authors: Muppalaneni, Naresh Babu; Prathima, Ch.; Sekhar, Akula Chandra
• Is Part Of: IOP conference series. Volume 1074:Issue 1(2021)
• Publication Date: 2021-02
• Subjects: Materials science -- Periodicals; 620.1105
• DOI: 10.1088/1757-899X/1074/1/012009
• Languages: English
• ISSNs: 1757-8981

Abstract

Abstract: In this present fast growing environment the automatic coin reorganization and identification machines has a vital role in all financial allied fields. At present most of coin recognition techniques are depends on physical properties of the coin like length, width, weight etc. Whereas image processing techniques are based on extraction of colour of the coin, edge features of the coin and shape of the coin. For recognition and detection of Brazil currency we have designed Machine Learning, Deep Learning (DL) models in this paper. We have designed various a Machine Learning (ML) models with Convolutional Neural Network for identification and reorganization of Brazil currency. Brazil currency consisting of 5 Centavos, 10 Centavos, 25 Centavos etc. Each of which has different shapes and designs. Different deep learning models are designed and applied ensemble to find the better accuracy. The trained ensemble model is tested on various the datasets which consists of shifting of images, rotation and translated images.