An Integrated Framework for Demographic Analysis, Person Recognition, and Path Tracking in Surveillance Systems

Public safety brings up a critical need for smart surveillance systems that go beyond the methodology conventions by providing a more in-depth, data-driven approach. This study proposes an important framework for person identification, demographic prediction, and path tracking to enhance situational awareness in monitored areas. Face recognition and gait analysis using MobileNet and VGG16 models have reached up to 94.8% and 98.5%, respectively, while predictions of age and gender are performed to provide demographic insights. The combination of YOLOv5 and Kalman filtering improved the path tracking accuracy from 64.68% to 88.45%, dealing with ID consistency and smoothing trajectories during occlusions, with an IoU of 0.86. The visual insights are enriched further using heatmaps to show high-activity zones and popular routes. Contextual path annotation with directional arrows and color-coded paths affords intuitive summarization that allows actionable insight into movement patterns and congregation areas. This will be a considerable stride forward toward a data-driven approach in surveillance for public safety.