This code is a pipeline for semantic segmentation which is based on pytorch-unet and tested on synthetic images/masks.
The pipeline consists of five parts:
- data preparation
- model creation
- model training
- data prediction
- index evaluation
Entry: main.py
- real data
- index evaluation
- cross validation