Thera: Aliasing-Free Arbitrary-Scale Super-Resolution with Neural Heat Fields

Alexander Becker^❄️🔥, Rodrigo Daudt^❄️🔥, Dominik Narnhofer^🔥, Torben Peters^🔥, Nando Metzger^🔥, Jan Dirk Wegner^🌶️, Konrad Schindler<sup>🔥

❄️</sup> Equal contribution
^🔥 Photogrammetry and Remote Sensing, ETH Zurich
^🌶️ Department of Mathematical Modeling and Machine Learning, University of Zurich

Thera is the first arbitrary-scale super-resolution method with a built-in physical observation model.

News

2025-03-15: We are #1 on Hacker News 🎉
2025-03-14: Interactive Hugging Face Space is online
2025-03-12: Pre-trained checkpoints are released

Setup environment

You need a Python 3.10 environment (e.g., installed via conda) on Linux as well as an NVIDIA GPU. Then install packages via pip:

> pip install --upgrade pip
> pip install -r requirements.txt

Use with pre-trained models

Download checkpoints:

Backbone	Variant	Download
EDSR-base	Air	Hugging Face | Google Drive
	Plus	Hugging Face | Google Drive
	Pro	Hugging Face | Google Drive
RDN	Air	Hugging Face | Google Drive
	Plus	Hugging Face | Google Drive
	Pro	Hugging Face | Google Drive

Super-resolve any image with:

> ./super_resolve.py IN_FILE OUT_FILE --scale 3.14 --checkpoint thera-rdn-pro.pkl

You can evaluate the models on datasets using the run_eval.py script, e.g.:

> python run_eval.py --checkpoint thera-rdn-pro.pkl --data-dir path_to_data_parent_folder --eval-sets data_folder_1 data_folder_2 ...

You can run python run_eval.py -h to display all testing options.

Training

Training code will be released soon.

Useful XLA flags

• Disable pre-allocation of entire VRAM: XLA_PYTHON_CLIENT_PREALLOCATE=false
• Disable jitting for debugging: JAX_DISABLE_JIT=1

Citation

If you found our work helpful, consider citing our paper 😊:

@article{becker2025thera,
  title={Thera: Aliasing-Free Arbitrary-Scale Super-Resolution with Neural Heat Fields},
  author={Becker, Alexander and Daudt, Rodrigo Caye and Narnhofer, Dominik and Peters, Torben and Metzger, Nando and Wegner, Jan Dirk and Schindler, Konrad},
  journal={arXiv preprint arXiv:2311.17643},
  year={2025}
}