Context-driven Abnormal Semantic Event Recognition for Healthcare Applications

Healthcare applications in a smart environment present an increasing need for technological support to monitor and recognize patients' activities in the hospital environment and their daily routine. Recognition systems for activities use knowledge models to define patients' routine activities, and deviations from the model are considered abnormalities. It models provide representation and reasoning, inferring implicit facts to discover high-level activities, analyzing and correlating events. We propose a knowledge-based hybrid reasoning approach to allow the recognition of abnormal semantic events. This approach consists of two stages: 1) modeling, ontology-based and probabilistic, supporting aspects with temporal and uncertainty data, 2) semantic reasoning, events are interpreted by a strategy context-driven hierarchical.