Prosthetic Communication software for individuals with special needs
My goal is to develop new approaches in speech technologies for physically or cognitively disadvantaged users. This includes applying the specialized automatic speech recognition (ASR) algorithms developed during my doctoral research into real-world tools for speakers with speech disorders. For example, I will develop software-assisted human-human interaction in which speech that is unintelligible because of physical disability is modified to produce a more comprehensible equivalent. This will involve acoustic transformations and speech recognition preprocessing. Much of my postdoctoral research will involve applying my expertise in ASR for atypical speech to Professor Mihailidis' COACH project. This project provides technologies that intelligently support older adults with special needs (e.g., dementia) in their home environment. I will integrate specialized ASR software into COACH that will interact in a dialogue with older adults and respond to emergency situations. This software will be developed at Quillsoft Ltd. and will be integrated into their flagship product, SpeakQ.