Status: published
With the increasing elderly population and a preference for independent living, there is a growing need for non-intrusive monitoring systems to ensure the safety and well-being of elderly individuals living alone. This paper presents a vision-based monitoring system that leverages real-time video streams and deep learning algorithms to detect unusual activities such as falls or prolonged inactivity. The system utilizes strategically placed cameras to capture video data, which is then processed and analyzed for anomalies. When abnormal behavior is detected, alerts are sent to caregivers via a Telegram API, enabling timely interventions. The system ensures privacy and security through encrypted data storage, while promoting independence and improving the quality of life for elderly individuals. By providing real-time monitoring and customizable alerts, the system also aims to reduce healthcare costs by facilitating early detection of potential health risks. The proposed solution demonstrates significant potential in enhancing elderly care, offering peace of mind to caregivers and fostering autonomy for the elderly.
Keywords: Vision-based monitoring, elderly care, fall detection, activity recognition, deep learning, real-time video surveillance, health monitoring, privacy protection, anomaly detection, smart homecare