In a bid to assist visually impaired people to easily determine denomination of currency notes, the Indian Institute of Technology at Ropar in Punjab has developed an Android App “Roshni”, using image processing and analytics.
The App was launched on Wednesday at the IIT Ropar. Assistant professor Puneet Goyal, said that ‘Roshni’ is the first Android App that works successfully with new currency notes that were released after demonetisation and it works well in broad range of light conditions and holding angles.
“This App utilises an adaptable deep learning framework, which further uses the patterns and features embedded on the notes to differentiate and determine the currency denomination. Our IPSA (Image processing, Security and Analytics) lab prepared a rich dataset of more than 13, 000 images of currency notes under varied real-world conditions. The user has to bring the currency note in front of phone camera and the App would provide audio notification intimating the currency note denomination to the user," said Mr. Goyal.
The team includes Dr. Puneet Goyal, Ph.D scholars Mandhatya Singh and Joohi Chauhan, and student R. Ram of the CSE & Engineering department.
The Seeing AI App by Microsoft is the only other App that facilitates recognition of both old and new Indian currency notes, but it is an iOS App and not available for school for visually impaired recently and performed the App testing there.
Ms. Nirmala, faculty member of the school, in a statement added “Earlier the visually impaired persons, differentiated the currencies based on the length and width of the note. But after demonetisation it became very difficult for them to identify the denomination, due to similarity in the sizes of new and old notes. Also, for those individuals who lost their vision due to old age or some other conditions, 'Roshni' will be very beneficial.”