This is the official implementation of Early Stopping for Deep Image Prior, which has been accepted to the TMLR. You can find our paper via either OpenReview or arXiv.
conda env create -f environment.yml
conda activate early-stopping
Google Colab contains all the dependencies our algorithms need. Thus, you are able to run our code directly on Google Colab. Moreover, we recommend you to access the dataset above via Google Drive. After uploading the dataset, you can run DIP with ES-WMV in ES-WMV.ipynb and DIP with ES-EMV in ES-EMV.ipynb, respectively.
We provide a ready-to-use dataset under the folder /Dataset where there are 4 types of noises we used in our paper where "XXX_2" indicates "low noise level", "XXX_3" indicates "medium noise level", and "XXX_4" indicates "high noise level".
Alternatively, you can also create the dataset by yourself. The clean images can be downloaded here. After you have the clean images, you can follow the ImageNet-C protocol (or write corruption functions by yourself) to create the corrupted images. For the parameters for each noise level, please check the Appendix of our paper.
@misc{wang2021early,
title={Early Stopping for Deep Image Prior},
author={Hengkang Wang and Taihui Li and Zhong Zhuang and Tiancong Chen and Hengyue Liang and Ju Sun},
year={2021},
eprint={2112.06074},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
- Hengkang Wang, wang9881@umn.edu, https://scholar.google.com/citations?hl=en&user=APqDZvUAAAAJ
- Taihui Li, lixx5027@umn.edu, https://taihui.github.io/
- Zhong Zhuang, zhuan143@umn.edu, https://scholar.google.com/citations?user=rGGxUQEAAAAJ
- Tiancong Chen, chen6271@umn.edu, https://sites.google.com/view/tiancong-chen
- Hengyue Liang, liang656@umn.edu, https://hengyuel.github.io/
- Ju Sun, jusun@umn.edu, https://sunju.org/