Exercise 3: Semantic Segmentation

📝 In this exercise, you will:

1. Prepare the U-Net baseline model and validation dataset.
2. Implement and use the Dice coefficient as an evaluation metric for the baseline model.
3. Improve metrics by experimenting with: Data augmentations Loss functions (bonus) Group Normalization, U-Net architecture

Setup

Please run the setup script to create the environment for this exercise and download data.

source setup.sh

Then run the exercise.ipynb notebook in your preferred tool.