Inspiration

We want to help visually impaired people in countries lacking modern and necessary infrastructure to accommodate their needs. We found out that it's not only 3rd world countries that might benefit from additional aid, but also large cities like New York with a large population of visually impaired that could certainly use a safety and quality of life upgrade.

What it does

Uses machine vision to analyze the environment and gives the user the needed feedback to safely cross intersections. Notifies the user when they get close to an intersection. If there are no traffic lights, checks for cars.

How we built it

Frontend: Webapp using HTML, CSS, JavaScript and Jquery Backend: Python (flask and geocoder), Azure Computer Vision

Challenges we ran into

Browser push notifications, accurate geo location

Accomplishments that we're proud of

Learning and implementing a new framework, linking of front- and backend

What we learned

New framework (flask), improve our knowledge of REST-APIs

What's next for TrafficSight

Develop native apps for iOS and Android, improve functionality, expand use of computer vision beyond traffic lights, migrate to a server

Share this project:

Updates