Home

Awesome

SeemoRe - ICML 2024

See More Details: Efficient Image Super-Resolution by Experts Mining

Eduard Zamfir<sup>1</sup>, Zongwei Wu<sup>1*</sup>, Nancy Mehta<sup>1</sup>, Yulun Zhang<sup>2,3*</sup> and Radu Timofte<sup>1</sup>

<sup>1</sup> University of Würzburg, Germany - <sup>2</sup> Shanghai Jiao Tong University, China - <sup>3</sup> ETH Zürich, Switzerland

<sup>*</sup> Corresponding authors

paper project demo

Latest

Method:

<br> <details> <summary> <font size="+1">Abstract</font> </summary> Reconstructing high-resolution (HR) images from low-resolution (LR) inputs poses a significant challenge in image super-resolution (SR). While recent approaches have demonstrated the efficacy of intricate operations customized for various objectives, the straightforward stacking of these disparate operations can result in a substantial computational burden, hampering their practical utility. In response, we introduce **S**eemo**R**e, an efficient SR model employing expert mining. Our approach strategically incorporates experts at different levels, adopting a collaborative methodology. At the macro scale, our experts address rank-wise and spatial-wise informative features, providing a holistic understanding. Subsequently, the model delves into the subtleties of rank choice by leveraging a mixture of low-rank experts. By tapping into experts specialized in distinct key factors crucial for accurate SR, our model excels in uncovering intricate intra-feature details. This collaborative approach is reminiscent of the concept of **see more**, allowing our model to achieve an optimal performance with minimal computational costs in efficient settings. </details>

Mixture of Low Rank Experts:

<p align="center"> <img src="figures/more.svg" width="600"/> </p>

Results:

<details> <summary> <font size="+1">Main Results</font> </summary> <p align="center"> <img src = "figures/main_table.png"> </p> </details> <details> <summary> <font size="+1">Visual Comparison</font> </summary>
HRBicubicSwinIR-LightDAT-LightSeemoRe (ours)
<img src="figures/img_1_HR_x4.png" height=90><img src="figures/img_1_Bicubic_x4.png" height=90><img src="figures/img_1_SwinIR_x4.png" height=90><img src="figures/img_1_DAT_x4.png" height=90><img src="figures/img_1_SeemoRe_x4.png" height=90>
<img src="figures/img_2_HR_x4.png" height=90><img src="figures/img_2_Bicubic_x4.png" height=90><img src="figures/img_2_SwinIR_x4.png" height=90><img src="figures/img_2_DAT_x4.png" height=90><img src="figures/img_2_SeemoRe_x4.png" height=90>
<img src="figures/img_3_HR_x4.png" height=90><img src="figures/img_3_Bicubic_x4.png" height=90><img src="figures/img_3_SwinIR_x4.png" height=90><img src="figures/img_3_DAT_x4.png" height=90><img src="figures/img_3_SeemoRe_x4.png" height=90>
<img src="figures/img_4_HR_x4.png" height=90><img src="figures/img_4_Bicubic_x4.png" height=90><img src="figures/img_4_SwinIR_x4.png" height=90><img src="figures/img_4_DAT_x4.png" height=90><img src="figures/img_4_SeemoRe_x4.png" height=90>
</details>

Install

Create a conda enviroment:

ENV_NAME="seemore"
conda create -n $ENV_NAME python=3.10
conda activate $ENV_NAME

Run following script to install the dependencies:

bash install.sh

Usage

Pre-trained checkpoints and visual results can be downloaded here. Place the checkpoints in checkpoints/.

In options you can find the corresponding config files for reproducing our experiments.

Testing

For testing the pre-trained checkpoints please use following commands. Replace [TEST OPT YML] with the path to the corresponding option file.

python basicsr/test.py -opt [TEST OPT YML]
Training

For single-GPU training use the following commands. Replace [TRAIN OPT YML] with the path to the corresponding option file.

torchrun --nproc_per_node=1 --master_port=4321 basicsr/train.py -opt [TRAIN OPT YML] --launcher pytorch

Citation

If you find our work helpful, please consider citing the following paper and/or ⭐ the repo.

@inproceedings{zamfir2024details,
  title={See More Details: Efficient Image Super-Resolution by Experts Mining}, 
  author={Eduard Zamfir and Zongwei Wu and Nancy Mehta and Yulun Zhang and Radu Timofte},
  booktitle={International Conference on Machine Learning},
  year={2024},
  organization={PMLR}
}

Acknowledgements

This code is built on BasicSR.