ICON (Inverse Consistency for Image Registration) is a powerful library for performing image registration using deep learning techniques. It implements the method described in the paper "ICON: Learning Regular Maps through Inverse Consistency" (ICCV 2021).
Key features:
- Deep learning-based image registration
- Regularization through inverse consistency
- Pre-trained models for knee and lung CT scans
- Easy-to-use PyPI package
This project is a fork of the original ICON repository, allowing for customization and experimentation with the framework.