# SAS-Net: Cross-Domain Image Registration as Inverse Rendering via Structure-Appearance Factorization

arXiv: https://arxiv.org/abs/2602.09050 · HTML: https://mlsanity.com/paper/2602.09050/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 25 references that are in the map.
- First posted on arXiv 6 Feb 2026.

## Builds on (most cited first, 11 of 11)

- [EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification](https://mlsanity.com/paper/1709.00029/) (2017)
- [Multimodal Unsupervised Image-to-Image Translation](https://mlsanity.com/paper/1804.04732/) (2018)
- [VoxelMorph: A Learning Framework for Deformable Medical Image Registration](https://mlsanity.com/paper/1809.05231/) (2018)
- [Diverse Image-to-Image Translation via Disentangled Representations](https://mlsanity.com/paper/1808.00948/) (2018)
- [SynthMorph: learning contrast-invariant registration without acquired images](https://mlsanity.com/paper/2004.10282/) (2020)
- [A survey on deep learning in medical image registration: new technologies, uncertainty, evaluation metrics, and beyond](https://mlsanity.com/paper/2307.15615/) (2023)
- [Alternative Telescopic Displacement: An Efficient Multimodal Alignment Method](https://mlsanity.com/paper/2306.16950/) (2023)
- [Zoom and Shift are All You Need](https://mlsanity.com/paper/2406.08866/) (2024)
- [DUAL: Dynamic Uncertainty-Aware Learning](https://mlsanity.com/paper/2506.03158/) (2025)
- [Learning Domain-Invariant Representations for Cross-Domain Image Registration via Scene-Appearance Disentanglement](https://mlsanity.com/paper/2601.08875/) (2026)
- [DCER: Dual-Stage Compression and Energy-Based Reconstruction](https://mlsanity.com/paper/2602.04904/) (2026)

## Abstract (arXiv)

Cross-domain image registration requires aligning images acquired under heterogeneous imaging physics, where the classical brightness constancy assumption is fundamentally violated. We formulate this problem through an image formation model I = R(s, a) + epsilon, where each observation is generated by a rendering function R acting on domain-invariant scene structure s and domain-specific appearance statistics a. Registration then reduces to an inverse rendering problem: given observations from two domains, recover the shared structure and re-render it under the target appearance to obtain the registered output. We instantiate this framework as SAS-Net (Scene-Appearance Separation Network), where instance normalization implements the structure-appearance decomposition and Adaptive Instance Normalization (AdaIN) realizes the differentiable forward renderer. A scene consistency loss enforces geometric correspondence in the factorized latent space. Experiments on EuroSAT-Reg-256 (satellite remote sensing) and FIRE-Reg-256 (retinal fundus) demonstrate state-of-the-art performance across heterogeneous imaging domains. SAS-Net (3.35M parameters) achieves 89 FPS on an RTX 5090 GPU. Code: https://github.com/D-ST-Sword/SAS-Net.