Development of Smart Money Classifier & Location Tracker Application Based on Convolutional Neural Network (CNN) for The Visually Impaired

Abstract

The number of visually impaired individuals in Indonesia continues to grow each year, caused by both congenital factors and age-related conditions. This situation positions Indonesia as the country with the second-highest number of blind people globally after Ethiopia. To address this, the present research introduces an Android-based mobile application named Smart Money Classifier & Location Tracker, specifically designed to assist visually impaired individuals in independently recognizing Indonesian banknote denominations and tracking their real-time location. The system integrates Convolutional Neural Network (CNN) technology for banknote image classification using a smartphone camera, enhanced with text-to-speech to announce recognition results. Moreover, a GPS-based tracking feature enables family members to monitor the user’s position through a dedicated interface that records chronological location history. Testing involved worn and folded banknotes under varied lighting and distances. Results indicate recognition accuracy of 85–99% under optimal conditions. For location tracking, the system estimated positions with average accuracy of 5–15 meters in open areas and 15–25 meters in semi-enclosed spaces, with a data transmission delay of 2–4 seconds. By combining these features, the application improves independence, safety, and convenience for visually impaired individuals, supporting them in everyday life.