IRNeXt: Rethinking Convolutional Network Design for Image Restoration

Yuning Cui, Wenqi Ren, Sining Yang, Xiaochun Cao, Alois Knoll
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:6545-6564, 2023.

Abstract

We present IRNeXt, a simple yet effective convolutional network architecture for image restoration. Recently, Transformer models have dominated the field of image restoration due to the powerful ability of modeling long-range pixels interactions. In this paper, we excavate the potential of the convolutional neural network (CNN) and show that our CNN-based model can receive comparable or better performance than Transformer models with low computation overhead on several image restoration tasks. By re-examining the characteristics possessed by advanced image restoration algorithms, we discover several key factors leading to the performance improvement of restoration models. This motivates us to develop a novel network for image restoration based on cheap convolution operators. Comprehensive experiments demonstrate that IRNeXt delivers state-of-the-art performance among numerous datasets on a range of image restoration tasks with low computational complexity, including image dehazing, single-image defocus/motion deblurring, image deraining, and image desnowing. https://github.com/c-yn/IRNeXt.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-cui23d, title = {{IRN}e{X}t: Rethinking Convolutional Network Design for Image Restoration}, author = {Cui, Yuning and Ren, Wenqi and Yang, Sining and Cao, Xiaochun and Knoll, Alois}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {6545--6564}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/cui23d/cui23d.pdf}, url = {https://proceedings.mlr.press/v202/cui23d.html}, abstract = {We present IRNeXt, a simple yet effective convolutional network architecture for image restoration. Recently, Transformer models have dominated the field of image restoration due to the powerful ability of modeling long-range pixels interactions. In this paper, we excavate the potential of the convolutional neural network (CNN) and show that our CNN-based model can receive comparable or better performance than Transformer models with low computation overhead on several image restoration tasks. By re-examining the characteristics possessed by advanced image restoration algorithms, we discover several key factors leading to the performance improvement of restoration models. This motivates us to develop a novel network for image restoration based on cheap convolution operators. Comprehensive experiments demonstrate that IRNeXt delivers state-of-the-art performance among numerous datasets on a range of image restoration tasks with low computational complexity, including image dehazing, single-image defocus/motion deblurring, image deraining, and image desnowing. https://github.com/c-yn/IRNeXt.} }
Endnote
%0 Conference Paper %T IRNeXt: Rethinking Convolutional Network Design for Image Restoration %A Yuning Cui %A Wenqi Ren %A Sining Yang %A Xiaochun Cao %A Alois Knoll %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-cui23d %I PMLR %P 6545--6564 %U https://proceedings.mlr.press/v202/cui23d.html %V 202 %X We present IRNeXt, a simple yet effective convolutional network architecture for image restoration. Recently, Transformer models have dominated the field of image restoration due to the powerful ability of modeling long-range pixels interactions. In this paper, we excavate the potential of the convolutional neural network (CNN) and show that our CNN-based model can receive comparable or better performance than Transformer models with low computation overhead on several image restoration tasks. By re-examining the characteristics possessed by advanced image restoration algorithms, we discover several key factors leading to the performance improvement of restoration models. This motivates us to develop a novel network for image restoration based on cheap convolution operators. Comprehensive experiments demonstrate that IRNeXt delivers state-of-the-art performance among numerous datasets on a range of image restoration tasks with low computational complexity, including image dehazing, single-image defocus/motion deblurring, image deraining, and image desnowing. https://github.com/c-yn/IRNeXt.
APA
Cui, Y., Ren, W., Yang, S., Cao, X. & Knoll, A.. (2023). IRNeXt: Rethinking Convolutional Network Design for Image Restoration. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:6545-6564 Available from https://proceedings.mlr.press/v202/cui23d.html.

Related Material