Crowd Sense: A Real-Time System for Crowd Analytics

Real-time crowd monitoring has become increasingly important in public and commercial environments such as shopping malls, campuses, transportation hubs, and event venues. Traditional surveillance systems rely on manual observation of video footage and do not provide meaningful real-time analytics. This paper presents Crowd Sense, a real-time crowd analytics system that detects and counts people from live video streams and performs gender classification using lightweight deep-learning models. The system employs MobileNet-SSD for person detection and a gender classification model for identifying male and female individuals. A Flask-based web dashboard integrated with SocketIO is developed to display live video, people count, gender distribution, and analytics updates without page refresh. The proposed system is designed to run efficiently on standard CPU-based hardware, eliminating the need for expensive GPU resources. Experimental results demonstrate reliable real-time performance under normal lighting conditions, making the system cost-effective, scalable, and suitable for practical deployment in real-world environments.