Multi-Modal Medical Imaging Decision Support at the Edge for Rapid Triage

Rapid medical triage increasingly depends on the ability to integrate heterogeneous imaging data under strict time, bandwidth, and privacy constraints. Centralized cloud based decision pipelines struggle to meet latency and resilience require ments in emergency and resource constrained settings. This work presents an edge centric multi modal medical imaging decision support framework that fuses radiological, physiological, and contextual signals to enable timely and explainable triage decisions. By combining deep feature fusion, lightweight edge inference, and adaptive model orchestration, the proposed approach supports real time clinical prioritization while preserving data locality. Experimental analysis demonstrates improved responsiveness, robustness, and diagnostic confidence across multiple emergency care scenarios.