期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2021
卷号：9
期号：7
页码：8759-8766
DOI：10.15680/IJIRCCE.2021.0907159
语种：English
出版社：S&S Publications
摘要：In this paper we summarize the importance of camera- based computing technology. It can help visually impaired people to automatically detect banknotes. A good algorithm for detecting paper money for people who are blind or visually impaired should have the following characteristics: 1) 100% accuracy, and 2) resilience in a variety of situations in a different environment and emergence. Many of the existing currency recognition algorithms are limited to restricted circumstances. In this project we propose a deep learning approach based on high accuracy. This approach is most effective in collecting more information in the classroom and is more effective in dealing with partial closure and change of views. In addition, reading transfer tests show its effectiveness in handling image rotation, measurements and lighting changes. To ensure the durability and standardization of the proposed method, we have compiled a large database of financial papers in a variety of contexts including closures, dense background, rotation, and lighting changes, measurements and views. The proposed algorithm achieves a high degree of recognition for our challenging database.
关键词：Indian Bank Notes Recognition;Visually Impaired;Transmission Reading;CNN
