An Interactive Multimodal Health Monitoring System with Prospective Federated Learning Integration

Globally, the prevalence of persistent medical conditions is rising, emphasising the critical requirement for secure, personalised, and effortlessly accessible health monitoring systems. Conventional healthcare systems often rely on clinical visits and centralised data storage, which limits real-time self-monitoring and increases privacy concerns. To address this gap, we propose an interactive, privacy-aware health monitoring system that enables users to monitor vital health indicators such as body mass index (BMI), basal metabolic rate (BMR), calorie intake, cholesterol levels, blood sugar, and blood pressure. Our system is built on a user-friendly Java Swing platform. It lets users enter health data, calculate health indicators, and view results in clear graphs and reports. The system also provides interactive recommendations. Any abnormal health conditions can be detected through this system. Unlike traditional systems, our platform is designed with future integration of federated learning, which will enable decentralised model training while ensuring user data remains on local devices. At this stage, federated learning is not yet implemented; it remains a conceptual future direction for subsequent iterations of the system. This ensures a decentralised model training and protects user privacy by keeping information on local devices. Although clinical diagnostics and evaluations remain the standard, they are not designed for routine self-monitoring or individual preventive measures. The system was tested with representative datasets and successfully calculated and visualised six key health metrics. This leads to a reduction in a critical gap in long-term care and self-awareness, especially in under-resourced or isolated environments.