A fast and effective posterior sampling framework for microscopy image dehazing
We are currently preparing the official release of HazeMatching, including:
- 🧠 Training and inference code
- 📊 Benchmark datasets (Zebrafish, Organoids, Microtubules, etc.)
- 📈 Evaluation scripts (PSNR, LPIPS, posterior sampling analysis)
- 🧪 Reproducible experiments from the paper
👉 Stay tuned — the full release will be available very soon!
HazeMatching is a posterior sampling-based method for microscopy image dehazing. Unlike one-shot restoration models, it generates multiple plausible reconstructions, enabling better uncertainty quantification and downstream analysis.
- ⚡ Fast sampling: Orders of magnitude faster than diffusion models
- 🎯 High-quality reconstructions: Strong PSNR and LPIPS performance
- 🔁 Posterior sampling: Generate diverse outputs instead of a single estimate
- 🔬 Calibrated uncertainty quantification: Provides uncertainty estimates for downstream analysis
HazeMatching: Fast Posterior Sampling for Microscopy Image Dehazing https://arxiv.org/abs/2506.22397
If you use this work, please consider citing:
@inproceedings{ray2026hazematching,
title = {HazeMatching: Dehazing Light Microscopy Images with Guided Conditional Flow Matching},
author = {Ray, Anirban and Ashesh, Ashesh and Jug, Florian},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition - FINDINGS Track},
year = {2026}
}We thank Francesca Casagrande, Alessandra Fasciani, Jacopo Zasso, Ilaria Laface, Dario Ricca, and Eugenia Cammarota for their valuable contributions to this work. We also acknowledge the support of Talley Lambert (Harvard Medical School) and Vera Galinova in setting up the microsim pipeline and some baselines, as well as the entire Jug Group for insightful discussions. This work was supported by the European Union through the Horizon Europe program (IMAGINE project, grant agreement 101094250-IMAGINE and AI4Life project, grant agreement 101057970-AI4LIFE) and the generous core funding of Human Technopole.
⭐ Star this repo to stay updated!