Papers › IRNeXt: Rethinking Convolutional Network Design for Image Restoration

IRNeXt: Rethinking Convolutional Network Design for Image Restoration

24 Apr 2023ICML 2023 4archive 2025-07-28

Yuning Cui, Wenqi Ren, Sining Yang, Xiaochun Cao, Alois Knoll

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.

Paper PDF Code

Code

c-yn/IRNeXt pytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Deblurring Image Deblurring Image Dehazing Image Restoration Rain Removal

Results from the paper archive 2025-07-28

Task	Dataset	Model	Metric	Value	Rank at snapshot	Leaderboard	Report
Deblurring	RSBlur	IRNext	Average PSNR	34.08	#5 of 12	Archive leaderboard	report
Image Deblurring	GoPro	IRNeXt	PSNR	33.16	#27 of 55	Archive leaderboard	report
Image Deblurring	GoPro	IRNeXt	SSIM	0.962	#27 of 55	Archive leaderboard	report
Image Dehazing	SOTS Indoor	IRNeXt	PSNR	41.21	#11 of 34	Archive leaderboard	report
Image Dehazing	SOTS Indoor	IRNeXt	SSIM	0.996	#11 of 34	Archive leaderboard	report
Image Dehazing	SOTS Outdoor	IRNeXt	PSNR	39.18	#6 of 31	Archive leaderboard	report
Image Dehazing	SOTS Outdoor	IRNeXt	SSIM	0.996	#6 of 31	Archive leaderboard	report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position Encodings Adam Attention BPE Convolution Dense Connections Dropout Label Smoothing Layer Normalization Linear Layer Multi-Head Attention Position-Wise Feed-Forward Layer Residual Connection Softmax Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections