README.md

---
license: mit
datasets:
  - Aldunitro/HDRT-TIR-diffusion-enhanced
language:
  - en
tags:
  - thermal
  - infrared
  - denoising
  - diffusion
  - transformer
task_categories:
- image-to-image
pretty_name: _Noise2DiffusionEnhanced_
size_categories:
- 10K<n<100K
---
&nbsp;

<div align="center">

<table>
<tr>

<td align="center" height="60">
<a href="https://openaccess.thecvf.com/content/CVPR2026W/PBVS/html/Hazebrouck_Self-supervised_Diffusion-guided_Hallucination-free_Thermal_Infrared_Image_Denoising_CVPRW_2026_paper.html">
<img src="./README_figures/button_paper.png" height="50"/>
</a>
</td>

<td align="center" height="60">
<a href="https://github.com/HensoldtOptronicsCV/Noise2DiffusionEnhanced">
<img src="./README_figures/button_github.png" height="50"/>
</a>
</td>

<td align="center" height="60">
<a href="https://huggingface.co/collections/SachyGuy/diffusion-guided-hallucination-free-tir-image-denoising">
<img src="./README_figures/button_huggingface.png" height="50"/>
</a>
</td>

</tr>
</table>

</div>

&nbsp;

---

# **_Noise2DiffusionEnhanced_ : A Pretrained Ready-to-use TIR Denoising Transformer**

**This is part of the official repository of the paper _Self-supervised Diffusion-guided Hallucination-free Thermal Infrared Image Denoising_, 2026.**  
&nbsp;

![Graphical Abstract](./README_figures/display_image.png)
_Our approach uses diffusion-based image enhancement and realistic TIR image degradation to generate image pairs for supervised learning (a) and leverages remarkable visual quality of diffusion models (c) without suffering from hallucinations (d-e)._  

&nbsp;
## Overview

The _Noise2DiffusionEnhanced_ is part of the work described in "Self-supervised Diffusion-guided Hallucination-free Thermal Infrared Image Denoising", F. Hazebrouck, A. Schock-Schmidtke, N. Stuhrmann,
J. Fottner, M. Teutsch (2026). The paper is available [here](https://openaccess.thecvf.com/content/CVPR2026W/PBVS/html/Hazebrouck_Self-supervised_Diffusion-guided_Hallucination-free_Thermal_Infrared_Image_Denoising_CVPRW_2026_paper.html).

The _Noise2DiffusionEnhanced_ is a pretrained [Uformer architecture](https://openaccess.thecvf.com/content/CVPR2022/html/Wang_Uformer_A_General_U-Shaped_Transformer_for_Image_Restoration_CVPR_2022_paper.html) [[1]](#1) for TIR sensor noise denoising. It was trained on the [HDRT-TIR-diffusion-enhanced](https://huggingface.co/datasets/Aldunitro/HDRT-TIR-diffusion-enhanced) dataset. The [pretrained weights](./model_best.pth) published here can be inserted in the [GitHub Project](https://github.com/HensoldtOptronicsCV/Noise2DiffusionEnhanced) for inference or fine-tuning.

This pretrained model should contribute to filling the current lack of reference denoising models for Thermal Infrared (TIR) single-image denoising, for direct use as well as for related research.


&nbsp;
## Files

- [model_best.pth](./model_best.pth) is the pretrained-weight file. 
It can only be used with the associated [project codebase](https://github.com/HensoldtOptronicsCV/Noise2DiffusionEnhanced), where is described how to insert it to run inference or fine-tune the weights.


&nbsp;
## Citation

If you use the _Noise2DiffusionEnhanced_, please cite our work:

    @article{hazebrouck_self-supervised_2026,
        title = {Self-supervised Diffusion-guided Hallucination-free Thermal Infrared Image Denoising},
        booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
        author = {Hazebrouck, Félix and Schock-Schmidtke, Alexander and Stuhrmann, Norbert and Fottner, Johannes and Teutsch, Michael},
        year = {2026},
        pages = {7091--7101},
    }


&nbsp;
## License and Copyrights  
The model weights were obtained using different code-compounds, each with its proper license:  
- The model architecture is a U-Former, introduced in [[1]](#1). The [original code](github.com/ZhendongWang6/Uformer) was adapted for our project and the modified version along with all license-notices with respect to the original source is available on our [project codebase](https://github.com/HensoldtOptronicsCV/Noise2DiffusionEnhanced) and is licensed under the MIT license. As we publish no code parts here, and because the code of [[1]](#1) only indirectly contributed to those weights, we cite them only for transparency.
- The model was trained with our _HDRT-TIR-DE_ dataset, which is licensed under the [CC-BY-NC-SA-4.0](https://spdx.org/licenses/CC-BY-NC-SA-4.0) license. The _HDRT-TIR-DE_ dataset itself is a modified version of the HDRT dataset introduced in [[2]](#2), but because the _HDRT-TIR-DE_ already comply with the licensing of the HDRT, as explicitly stated on the [_HDRT-TIR-DE_ repository](https://huggingface.co/datasets/SachyGuy/HDRT-TIR-DE) and only indirectly contributed to the weights, only the license from the _HDRT-TIR-DE_ dataset applies here.   
According to the license of the _HDRT-TIR-DE_ Dataset ([CC-BY-NC-SA-4.0](https://spdx.org/licenses/CC-BY-NC-SA-4.0)), any "adapted material" must be licensed under the same or a compatible license, therefore including these pre-trained weights.
- The degradation model used to synthesize the noisy counterparts for the clean images from the _HDRT-TIR-DE_ dataset for the supervised training of the UFormer is adapted from the noise model introduced in [[3]](#3). The [original code](https://github.com/cailijing/MDIVDnet) was adapted for our project and the modified version along with all license-notices with respect to the original source is available on our [project codebase](https://github.com/HensoldtOptronicsCV/Noise2DiffusionEnhanced) and is licensed under the [MIT license](https://opensource.org/license/mit). As we publish no code parts here, and because the code of [[3]](#3) only indirectly contributed to those weights, we cite them only for transparency.

Therefore, the pre-trained weights published in this repository inherit the [CC-BY-NC-SA-4.0](https://spdx.org/licenses/CC-BY-NC-SA-4.0) license from the _HDRT-TIR-DE_ dataset. According to this license, we include the original dataset authors, a copyright notice, a link to original dataset, a link to license, and a statement of modifications.

Authors:  
© Félix Hazebrouck, Alexander Schock-Schmidtke, Norbert Stuhrmann, Johannes Fottner, Michael Teutsch

Original dataset:  
_HDRT-TIR-DE_ dataset, available at https://huggingface.co/datasets/SachyGuy/HDRT-TIR-DE, introduced in _Self-supervised Diffusion-guided Hallucination-free Thermal Infrared Image Denoising_ by Hazebrouck et al., 2026.

License:  
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International  
https://spdx.org/licenses/CC-BY-NC-SA-4.0

Modifications:  
We encoded the statistical information from the _HDRT-TIR-DE_ dataset in the hereby published pretrained neural-network architecture weights through a process of machine learning.

These pretrained weights are distributed under the same license (CC-BY-NC-SA-4.0).


## References
<a id="1">[1]</a> 
Zhendong Wang, Xiaodong Cun, Jianmin Bao, Wengang Zhou, Jianzhuang Liu and Houqiang Li. _Uformer: A general u-shaped transformer for image restoration_. CVPR, 2022

<a id="2">[2]</a> 
Jingchao Peng, Thomas Bashford-Rogers, Francesco Banterle, Haitao Zhao and Kurt Debattista. _HDRT: A large-scale dataset for infrared-guided HDR imaging_. Elsevier Information Fusion, 120, 2025

<a id="3">[3]</a> 
Lijing Cai, Xiangyu Dong, Kailai Zhou and Xun Cao. _Exploring video denoising in thermal infrared imaging: Physics-inspired noise generator, dataset, and model_, IEEE 2022