Disability Rights
The app that tells blind users what the camera sees
Luke Corneil, an I-STEM student at Thomas A. Blakelock High School, is building NaviSight, an app that turns a phone camera into spoken descriptions of the surroundings for blind and visually impaired users, a group he says includes over 1.2 million Canadians.
The phone camera reads whatever is in front of the user, in real time. An AI system describes the scene and the app speaks it back. That is NaviSight, the accessibility app Luke Corneil is building so that blind and visually impaired users can navigate their surroundings more safely and independently.
Over 1.2 million Canadians have blindness or vision loss, a figure verified by the Canadian National Institute for the Blind, and they often need help from other people to take part in everyday tasks. The tools that exist for them, he says, are either insufficient or unaffordable. NaviSight plans to address this "with high functionality, accessibility features and low costs."
“I’ve had a lot of opportunities through I-STEM and Shad to explore ideas for how new technologies can make people’s lives better. Now, I want to take those ideas further and build them into something that can make a genuine difference.”
Luke Corneil
The first Android Studio prototype analysed individual images, describing surroundings, identifying differences between scenes and giving the user spoken feedback. It has since progressed to an Android application with real-time camera analysis and AI-powered scene description. It is still in development.
NaviSight began in July 2025 at the Shad summer program at the University of Guelph, where his group created it for a case study competition on transportation. That first version was for navigating train stations. It won the "Most Innovative Concept" award.
After the program he decided to develop it into an actual product, and the idea shifted from train stations to general world navigation. His first prototype was a Python script using a YoLo vision model to highlight key objects in an image. He entered it in the Samsung Solve for Tomorrow Competition, where it was a top 20 national submission. The Android Studio version came next.
There is a second project. He also built CubeSpeed, a web-based app for Rubik’s cube timing and training. A competitive speedcuber himself, he says the most popular apps and websites missed many of the features he wanted. He has published the site and shared it with others at a competition.
"For me, the most exciting part of an idea is turning it into something that actually works," he says. "Through NaviSight and CubeSpeed, I want to build things that solve real problems."