What it does
Iris is an affordable, effective, and accurate device that prevents drivers from distracted driving and drowsiness.
68% of collisions are caused by distracted driving. There's been a 14% of increase in traffic fatalities since 2014. Nearly 1.3 million people die in road crashes each year.
With the growing number of accidents and fatalities, how can everyone be safe on the road? What measures can we take to reach a safe road?
We at Iris set out to answer this question. We investigated car accidents and the current prevention devices available on the market, and found that prevention devices for distracted and drowsiness are not only large and bulky, but were expensive.
Realizing that everyone could not afford expensive monitoring systems, we set out to develop our own low-cost and accurate device to create a safe - road for you and those around you.
How we built it
We began our research with an article called "Fast and Accurate Algorithm for Eye Localization for Gaze Tracking in Low Resolution Images." We used the results from this study and others to detect signs of drowsiness and distraction. To implement Iris, we used Python, a Raspberry Pi + camera, Flask for the backend, Microsoft Azure for the Face Landmarks, and Twilio for the Voice Calls.
Challenges we ran into
The biggest challenge was using Computer Vision, a concept that was really challenging to implement. None of us on the team did anything remotely to Artificial Intelligence, so we had lows where we didn't believe in ourselves and our idea. We began with OpenCV, but ended up having troubles with downloading the dependencies such as dLib, so we switched to Microsoft Azure, a clean and simple interface to deploy apps with the FaceAPI.
Accomplishments that we're proud of
We're proud of not accepting defeat in the face of adversity. Around 11 PM on Saturday morning, we felt defeated by OpenCV and it's numerous challenges. We wanted to turn back and head home with faces of defeat and shame. But around 1 PM, after chatting to a Microsoft representative at the Expo, we learned that we could use Azure to deploy our Face Landmarks, a step that changed the course of the project and our motivation.
What we learned
We learned that Artificial Intelligence can be rewarding and impactful, despite its challenges. We also learned that the idea of hackathons is to bring a community of ambitious students who want to create a happier future.
What's next for Iris
We'd like to expand our data that displays driver statistics and bring it to an app in the Google and iOS Play Stores. We'd also like to find a simpler setup to deliver a user experience that doesn't obtrude the view of the road. Most importantly, we'd like to develop a community of drivers that are safe on the roads and beyond.