DashGuard

DashGuard is an AI-powered system developed as part of my capstone R&D project during my Master’s studies at NUS. It monitors customer exits and detects unpaid departures without relying on physical gates or alarm gantries. The system tracks individual and group movement using lightweight detection and re-identification models, while analyzing diner behavior to dynamically form groups based on seating, entry timing, and social interactions. It then maps group IDs to order IDs by integrating with POS systems or through gesture and action recognition at cashier zones. Finally, unpaid exits are flagged in real time using intelligent zone detection and behavioral logic, making DashGuard a seamless and efficient solution for modern dining environments.