Team Members: Sean Perry, Jenny Lam Jason Liang, Aran Punniamoorthy
This repo includes an implementation of a semantic segmentation model which predicts the semantic class label for each pixel in an image across all images in a cloud image dataset.
An end-to-end model training + inference notebook can be found in the src folder.
Our data comes from WSISEG (click here) and HYTA Databases (click here) to download our dataset.
You can then structure the data by putting it into a folder structure as such:
|__ src
|__ data
|__ train
|__ val
|__ test
We would like to thank Jonathan Zamora and Joseph Liu from ACM AI at UCSD for advising our team on this project as part of the ACM Projects Cohort for Fall 2021.