UWindsor computer science grad creates speech-to-text tool
'Anytime there’s a speaker and an audience, this could be applied'
Classroom learning may have just gotten easier for people with hearing impairments.
Computer science graduate Selina Gabriele has developed a distributed speech to text accessibility tool that she hopes replaces the volunteer student note taker program at the University of Windsor: Aida Note.
In real-time, students can log on to a website and see the instructor's words appear. Professors wear a microphone while they lecture to make it happen.
Gabriele said the inspiration for the AidaNote project came from her nephew Tanner.
"Tanner is a child who was born completely deaf and he's just started his first year of school," said Gabriele. "That really kind of sparked me to take my experience in writing accessibility tools and apply it to the speech to text realm."
According to Gabriele, only about 60 per cent of the need for note takers is filled by volunteers at the university.
Gabriele and partner Stevan Ljuljdurovic won three awards at the 2018 RBC EPIC Business Model Canvas Competition in November, taking home home more than $12,000 in prizes.
Speech-to-text transcription already exists in everyday platforms like OK Google and Apple's Siri. It also already exists in accessibility systems for students. What makes Gabriele's different is the distribution model.
"There's one speaker, but unlimited people can log in," said Gabriele. "What we're worried about is professors don't want a tool that make it so students don't have to go to class."
Due to that worry, Gabriele is looking at adding restrictions, such as needing to be on the same wireless network as the speaker, in order to access the website. She also hopes it has implications beyond assisting the hearing-impaired, and could be used for learning disabilities as well.
"Anytime there's a speaker and an audience, this could be applied," said Gabriele.
Gabriele says they're about halfway through the development process, focusing right now on refining the transcription tool. She hopes to have it completed by April 2019, for a pilot during the last month of school. Then the plan is to soft launch in the summer for free, followed by a paid program once school begins Fall 2019.
With files from Tony Doucette