- Light, shows you a way to navigate through problems. We wished to present a solution to tackle low light photography which is a major feature in development over recent years.
What it does
- See in the Dark
How we built it
- Using Fully Convolutional Neural Network on pytorch for the model, Flask/HTML/CSS for the web app.
Challenges we ran into
- We tried running our model natively on Android but couldn't do it due to image compatibility and ram issues. We converted it into a web app instead.
Accomplishments that we're proud of
- We have created a small and efficient neural network model compared to existing solutions. We created a cool UI we are proud of.
What we learned
- Modelling in PyTorch, HTML/CSS designing.
What's next for See in the Dark
- Making it run real time on mobile device.