An Intelligent Banknote Recognition System by using Machine Learning with Assistive Technology for Visually Impaired People

Hong Kong Dollar is the main currency used in Hong Kong. Hong Kong banknotes are issued by three commercial banks, namely Hongkong and Shanghai Banking Corporation (HSBC), Standard Chartered Bank, and Bank of China. People can typically differentiate the different denominations of coins and banknotes easily by their characteristics. For visually impaired, they can simply determine the values of coins by tactile sense due to the unique size and shape of each type of coin. However, it will be difficult for visually impaired to differentiate different denominations of different types of banknotes. Although the size of each banknote category is different, and it contains characteristics like high-tactility numeral at top left-hand corner, braille and tactile lines at the bottom left-hand corner of the banknote. These features can be used for recognition when the banknote is pretty new. When the banknote becomes old, the features will be unapparent, so visually impaired are hard to confirm the values of the banknotes. Moreover, cheating in purchasing will be much easier in paying the amounts or collecting the change during monetary transactions. In this paper, a tailor-made intelligent banknote recognition system is developed by applying machine learning with assistive technology for visually impaired to determine the denominations of the banknotes effectively and efficiently. The application can be used in a smartphone with sound and vibration to facilitate the daily monetary consumption of the visually impaired people.