This repo uses deep learning (U-Net, SegNet, DeepLabV3Plus) to automatically detect fractures from images.
Highlights:
- Configurable models (loss, architecture, hyperparameters)
- Preprocessed data (e.g., matteo21, ovaskainen24)
- Training & inference scripts (main.py, notebooks, shell scripts)
- Model code (data handling, loss functions, architectures, training)
- Trained model outputs (weights, predictions)
Get Started:
- Create a dev env and install dependencies
- Refer to scripts and configs for training and inference.
- Explore notebooks for interactive experimentation.