Paper money recognizer using feature descriptor

Abstract

People are still using paper money for daily transaction; this, however, will expose some difficulty for visually impaired people. Though they can still read the nominal value of the paper money by the help of other people or by touching the tactile feature, they cannot depend upon others all the time nor touching the tactile feature properly if the paper money is worn. Some alternatives have been proposed and conducted. One of them is using money value recognition application. The application will recognize nominal value of paper money comparing the image of the paper with database. This process is using a feature extraction algorithm called ORB feature descriptor. It has been used for six (6) different types of currencies that are 5 most traded currencies and Indonesia currency and also for different types of nominals (bills).