DISTANT TALKING SPEECH RECOGNITION IN SURGERY ROOM : THE DOMHOS PROJECT

In recent years, ASR technologies were successfully applied in many commercial fields such as web-search, dictation, car-control, automated voice answering system as well as radiological reporting. Most state of the art applications are still based on close-talking solutions, imposing the constraint to speak always very close to a microphone -equipped device. Although this could lead to better performance, in some contexts users would prefer to use the technology in an hands-free modality. The purpose of the DomHos project is to lead distant -talking speech recognition technologies inside the surgery room, a challenging but fascinating scenario. In particular, the first goal is to enable the surgeon to dictate some speech notes during the operation itself. This could be potentially very helpful since, at the end of the surgery, the doctor already has a draft report about the operation, avoiding the boring activity of writing it from scratch. The distant-talking note dictation procedure requires the design of both an online keyword-spotting and an off-line transcription system able to cope with high reverberation and very high non-stationary noises. The other goal regards the speech technology based implementation of the checklist protocol, a mandatory procedure proposed by the World Health Organization (WHO) and recently introduced also in Italy by the Health minister to prevent mistakes, mortality and complications during the surgery