Inspiration:

Girls or boys wear lipstick almost every day. Guys want to get a lipstick for his/her partner but have no idea which to get. Customers might want to try it on before they purchase the lipstick.

What it does:

Lipstick recognitions and virtual makeup.

How we built it:

Pytorch & Messenger API

Challenges we ran into:

No available data. We have to generate our own data.

Accomplishments that we're proud of:

Our team is able to achieve 96% accuracy on the test data.

What we learned: Messenger API.

Learn how to apply deep learning in Pytorch.

What's next for RECOLIP:

  1. Extend more categories for data-set
  2. Extension for cosmetics in general
  3. Product recommendations
Share this project:
×

Updates