Real-time acoustic detection on mobile devices to support the deaf and hard of hearing with situational awareness

Hearing loss individuals face apparent as well as subtle, systematic, and safety challenges every day in a world predicated on hearing. Traditional hearing assistive solutions are not readily available in some environments. Hence, Lisnen is exploring opportunities to leverage available tools, such as mobile devices, to act as a hearing assistive device to provide situational awareness. These devices will identify relevant and vital sounds, both indoors and outdoors spaces using deep learning models. However, running deep learning models, while continuously streaming audio input at the same time, increases energy consumption. For this project, we want to explore ways to improve the performance and design networks that reduce mobile GPU power consumption.

Intern: 
Amr Gaballah
Superviseur universitaire: 
Tiago H Falk
Province: 
Quebec
Partenaire: 
Partner University: 
Programme: