neural-

Inspiration

We wanted to create an interesting machine learning project that utilized models for improved accuracy.

What it does

It identifies a human image segment within a jpg or png image file and then extracts and reintegrates that image into a background that has been modified by the neural style model.

How we built it

We utilized the CRF-RNN Semantic Image Segmentation model to accurately identify human image segments within an image file.

Challenges we ran into

We had difficulty training the neural style model and getting all of the pre-built machine learning tools to function properly, but eventually we managed to solve most of these issues.

Accomplishments that we're proud of

What we learned

What's next for neural-cycle

Built With

• c++
• jupyter-notebook
• python