Deep Learning - Driver Edge Registration for Enhanced Home Health Monitoring

During pandemics like COVID-19, healthcare systems are overwhelmed by the rapid spread of contagious diseases, leading to limited resources. To ease this strain, home healthcare for non-critical patients and elderly individuals living alone offers a practical solution. Home health monitoring uses non-intrusive visual sensors for continuous tracking, reducing hospital loads. This study proposes a transfer learning-based edge computing method for home health monitoring, utilizing a pre-trained convolutional neural network (CNN), specifically the VGG16 model. This model is fine-tuned for deployment on AI-enabled sensors connected to edge cloud servers. The objective of this research is to develop a vision-based AI-powered patient monitoring system for the improvement of healthcare by identifying key health events such as falls, abnormal motion, and atypical breathing in real time. Compared to traditional wearable sensors, the system utilizes cameras and deep neural networks for contactless monitoring that is more precise and comfortable. Edge computing is integrated to ensure faster processing and real-time alerting to caregivers, reducing response time. Data privacy, reliability, and interfacing with healthcare systems are also addressed in the research. The results show that AI-powered vision-based monitoring enhances patient safety, reduces caregiver workload, and provides an economical, scalable solution for real-time health monitoring for hospitals and assisted living facilities