Overview
Rupiah Currency Denomination & Sum Detection is an assistive AI system designed to help visually impaired individuals identify and count Indonesian banknotes automatically using computer vision and YOLOv8-based object detection.
Problem to Solve
Build an accessible and reliable smart currency detection system that enables blind and visually impaired users to recognize Rupiah denominations and calculate total money independently in everyday situations.
Solution Approach
Led the project development and contributed to the AI pipeline by preparing and managing a dataset of 2,877 Indonesian Rupiah images captured across varied lighting conditions, camera angles, and real-world scenarios, then training and optimizing a YOLOv8 model for denomination detection. The system was designed to process camera input in real time, detect multiple bills in a scene, classify each denomination accurately, and estimate the total amount for practical assistive use.
Impact
Produced a functional assistive technology prototype with over 90% precision and recall in most testing conditions, including low light and moderate distortion, demonstrating strong potential for improving independence and daily financial accessibility for visually impaired users.