The SAFERS project will develop a set of innovative tools and techniques for performing real-time analysis of emergency calls. These tools and techniques will build on recent technological advances in speech processing, language technology and machine learning, with the goal of enhancing the quality, efficiency and safety of emergency responses. The project will need to address several important R&D challenges. One major research effort will be the development of a speech recognition system for Norwegian able to transcribe emergency calls with sufficient accuracy. The project will notably investigate how to integrate environmental noise, psychological stress and other emotional factors into the acoustic models of the speech recogniser. Dialogue modelling techniques will be applied to track in real-time the interactions between the caller and the emergency operator and automatically extract important pieces of (medical) information from the dialogues. Finally, the project will apply machine learning techniques to automatically predict relevant variables (such as the situation's urgency level or the most likely geographical position of the caller) on the basis of the information collected during the call, combined with various contextual factors. These predictions will be subsequently exploited to quickly detect potential errors, omissions or deviations from operational guidelines. The SAFERS project brings together an international consortium of leading researchers in the fields of speech recognition, natural language processing, statistical modelling and machine learning. The project's stakeholders (represented by the National Centre on Emergency Communication in Health) will also take an active role in the SAFERS project, both regarding the design of technological solutions that are best suited to the practical needs of emergency response services, and the empirical evaluation of these solutions in real-life scenarios.
Project leader: Pierre Lison