TraLiAI

Inspiration

We believe in using technology to make the world a safer place.

What it does

TraLi AI is a faster R-CNN developed to recognize and classify traffic lights.

How we built it

We used a faster R-CNN implementation and built upon it, training the network on the Small Bosch Traffic Light Dataset.

Challenges we ran into

The implementation of the network and using the Bosch Dataset for training purposes. Computing resources and internet connectivity also proved to be very troublesome.

Accomplishments that we're proud of

Unwillingness to give up and a joint team effort to achieve the best in a very limited time.

What we learned

How to actually implement a CNN, fine tune it and adjust it to work with the given dataset as part of the challenge.

What's next for TraLiAI

Change the world...by further optimizing and improving the network.

Built With

• keras
• python