For this project I will be using the fast.ai framework and the Inria Aerial Image Labeling Dataset to train a semantic segmentation model. The goal is to use this model to run inferences on new imagery and to expose these inferences on an interactive map inspired by daveluo.
Current progress on the model can be found in the jupyter notebook