FeeTap: A Smart Payment Kiosk with Facial Recognition for Automated Student Department Fee Collection

Traditional manual fee collection at academic institutions often leads to administrative bottlenecks and prolonged wait times. To address these inefficiencies, this study developed and evaluated FeeTap, an automated self-service payment kiosk prototype integrated with biometric facial verification. Implemented using a Raspberry Pi 4B architecture and Python’s face_recognition library, FeeTap features a multi-denominational currency acceptor, automated change dispensing, and a dedicated mobile application for real-time transaction monitoring. Utilizing a quasi experimental design, the researchers conducted preliminary performance testing with a purposive sample of computer engineering students to assess biometric accuracy and operational throughput. Results indicated that the facial recognition module achieved a 72.92% cumulative success rate within a three-attempt authentication protocol. The remaining 27.08% composite failure rate consisted of both recognition timeouts and misidentifications, suggesting sensitivity to environmental micro-variations. Furthermore, a paired-samples t-test of successfully authenticated participants (n = 35) revealed a statistically significant difference between the FeeTap and manual methods in transaction times (t(34) = 12.59, p < .001). For this subgroup, the FeeTap system demonstrated a 44.77% reduction in mean transaction time, from 121.06 seconds to 66.86 seconds, including recognition latency. While a strong positive correlation (r = .77) was observed, the high failure rate and preliminary accuracy suggest that the system currently functions as a proof of-concept. These findings indicate that while FeeTap offers a promising alternative to manual transactions, further optimization of the biometric layer and failure-handling protocols is required for campus-wide deployment.