# AIRNet: Self-Supervised Affine Registration for 3D Medical Images using Neural Networks

arXiv: https://arxiv.org/abs/1810.02583 · HTML: https://mlsanity.com/paper/1810.02583/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Oct 2018.

## Built on by (most cited first, 12 of 12)

- [Equivariant Filters for Efficient Tracking in 3D Imaging](https://mlsanity.com/paper/2103.10255/) (2021)
- [Test-Time Training for Deformable Multi-Scale Image Registration](https://mlsanity.com/paper/2103.13578/) (2021)
- [RobustGS: Unified Boosting of Feedforward 3D Gaussian Splatting under Low-Quality Conditions](https://mlsanity.com/paper/2508.03077/) (2025)
- [SE(3)-Equivariant and Noise-Invariant 3D Rigid Motion Tracking in Brain MRI](https://mlsanity.com/paper/2312.13534/) (2023)
- [Closed-Loop LLM Discovery of Non-Standard Channel Priors in Vision Models](https://mlsanity.com/paper/2601.08517/) (2026)
- [Medical Imaging with Deep Learning for COVID- 19 Diagnosis: A Comprehensive Review](https://mlsanity.com/paper/2107.09602/) (2021)
- [Medical Image Registration and Its Application in Retinal Images: A Review](https://mlsanity.com/paper/2403.16502/) (2024)
- [LEMUR 2: Unlocking Neural Network Diversity for AI](https://mlsanity.com/paper/2607.06839/) (2026)
- [D-Net: Siamese based Network with Mutual Attention for Volume Alignment](https://mlsanity.com/paper/2101.10248/) (2021)
- [Application-driven Validation of Posteriors in Inverse Problems](https://mlsanity.com/paper/2309.09764/) (2023)
- [Decoupled Geometric Parameterization and its Application in Deep Homography Estimation](https://mlsanity.com/paper/2505.16599/) (2025)
- [Scaling Closed-Loop Feature Channel Configuration with LLMs](https://mlsanity.com/paper/2607.20516/) (2026)

## Abstract (arXiv)

In this work, we propose a self-supervised learning method for affine image registration on 3D medical images. Unlike optimisation-based methods, our affine image registration network (AIRNet) is designed to directly estimate the transformation parameters between two input images without using any metric, which represents the quality of the registration, as the optimising function. But since it is costly to manually identify the transformation parameters between any two images, we leverage the abundance of cheap unlabelled data to generate a synthetic dataset for the training of the model. Additionally, the structure of AIRNet enables us to learn the discriminative features of the images which are useful for registration purpose. Our proposed method was evaluated on magnetic resonance images of the axial view of human brain and compared with the performance of a conventional image registration method. Experiments demonstrate that our approach achieves better overall performance on registration of images from different patients and modalities with 100x speed-up in execution time.