The scientific scope of CASPR is within statistical signal processing, machine learning, information and communication theory with applications to wireless exchange of information between listening devices and other external devices, pattern recognition and data mining in body worn sensor data, and perception-based statistical signal processing.
With recent rapid advances in machine learning (ML) and artificial intelligence (AI), miniaturization of low-power transducers other than microphones, as well as power-efficient chipsets specifically tailored towards ML/AI solutions, we see a wealth of possibilities for groundbreaking signal processing research to benefit hearing impaired end-users. In order to make best use of these technological advances, the research scope of CASPR-II will be spanned by the following three thematic pillars:
- Multiple Modalities: Solutions that rely on an extended set of input sensors in addition to microphones, for example, accelerometers, gyros, body-worn electrodes such as EEG sensors, and even cameras. Additional modalities could include vision, e.g. information presented via a screen of sorts.
- User-Symbiotic Solutions: Solutions that aim at cooperating with the end-user will – to a much larger extent – be able to present the acoustic signal-of-interest to the end-user at any given moment in time.
- Beyond Audibility: Exploring signal processing solutions – primarily fueled by recent advances in ML-based speech conversion techniques – to do more than simply restoring audibility.