Inspiration

Our team was inspired to find a solution for safer driving after considering the recent legalization of marijuana. Being impaired due to being high, drunk or tired is a very pertinent issue. Impaired driving not only puts the driver at risk but also risks the lives of everyone else on the road. If a driver was to be impaired and remain unresponsive behind the wheel, they would most likely cause an accident on the road.

What it does

Blink uses Microsoft's custom vision AI trained with photos of open and closed eyes. With a webcam, companies such as Uber, Lyft and Taxis can have service which monitor's their drivers (specifically whether they become impaired/unresponsive) if they drive a smart car. When Blink detects that the user's eyes have been closed for 5s, the system issues an alert using Twilio. Immediately the latitude and longitude are forwarded to an emergency contact of the driver, telling them the ID of the car in distress and the latitude and longitude of the car.

How we built it

We are using Python and flask as the backend and angularjs as the front end. We used Azure computevision to train a model on open and closed eyes and then used that model to detect if the driver has fallen asleep by computing the amount of time their eyes are closed. We build a flask rest API using which the frontend predicts whether the driver has closed their eyes. we are using frames from the video captured by the webcam and detect closed eyes. Then we use the smartcar API to find the location of the smartcar and using twilio's SMS API we notify the driver's first responders and even the nearest police station. The frontend is developed by javascript with angular to parse json files from a jsonstored server with the info about the status of the driver, The status is displayed on the webpage by DOM manipulation.

Challenges we ran into

We ran into problems trying to parse the JSON files from the servers, we also ran into problems trying to set up automated HTTP get and push requests to the Jstore server.

Accomplishments that we're proud of

We're proud of what we were able to produce with the time and resources that we had, for most of us this was our first hackathon. We're proud of being able to successfully train the neural network to recognize faces with closed or open eyes and being able to communicate this information with the front-end and back-end seemlessly.

What we learned

We learned how to integrate various APIs and try many different ways to achieve the task and choose the best and the most efficient way to do so.

What's next for Blink

We would like to expand to notifying the nearest police department in relation to the GPS of the smart car. The police can then be notified that there is a potentially distracted driver in their area and get updates as to the location of the car. This would be done using Google's firebase and a few lines of code determining how to find the nearest police station to the current GPS.

Share this project:
×

Updates