Fall Detection and Alert System for Elderly People

Ensuring the safety of elderly individuals has become increasingly important as mobility and health challenges grow with age. Human Activity Recognition (HAR) contributes greatly to this need by monitoring routine actions such as sitting and standing, while also detecting critical events like falls. Since falls are a major cause of injury among older adults, developing a dependable and automated detection method is essential. This project presents a system designed to analyze human posture and identify activities in real time, supported by an easy-to-use interface for continuous observation.The system processes video input either from a live camera or an uploaded file by examining each frame to track changes in body movement. A classification model is applied to differentiate between actions such as standing, sitting, and falling based on posture and motion patterns. When a fall is recognized, the system immediately triggers an alert to ensure timely support. The interface allows users to monitor activities smoothly and access results in real time.The outcomes show that the system can accurately detect key activities and maintain consistent performance across different video sources. The quick fall-detection response further demonstrates its usefulness in safety monitoring. Overall, the system offers a practical, reliable solution for elderly activity recognition and shows strong potential for future improvements in assistive healthcare technologies.