Authors - ES Sithpahan, NH Wanigasingha, MKA Ariyaratne, PRS De Silva Abstract - This study presents the design and implementation of an interactive coin classification and saving assistant aimed at enhancing financial literacy in children. The system combines computer vision and machine learning techniques to automatically identify Sri Lankan coins using a custom image dataset. A convolutional neural network (CNN) was developed and evaluated using key performance metrics, including accuracy, precision, recall, F1-score, and confusion matrices, to ensure robustness and reliability. The software was integrated with hardware components comprising a Raspberry Pi, touchscreen interface, servo motor, and webcam, forming a tangible coin-till device for user interaction. The end-to-end system, from coin insertion and classification to actuation and feedback, was validated through both quantitative and qualitative testing. Quantitative evaluation focused on model performance, while qualitative analysis assessed usability, engagement, and educational effectiveness based on feedback from children and parents. Results indicate high usability and engagement, with participants demonstrating increased interest in saving behavior. The study highlights the feasibility of combining AI and embedded systems to deliver educational experiences for children.
