Papers › Revitalizing Convolutional Network for Image Restoration

Revitalizing Convolutional Network for Image Restoration

25 Jun 2024IEEE Transactions on Pattern Analysis and Machine Intelligence 2024 6archive 2025-07-28

Yuning Cui, Wenqi Ren, Xiaochun Cao, Alois Knoll

Image restoration aims to reconstruct a high-quality image from its corrupted version, playing essential roles in many scenarios. Recent years have witnessed a paradigm shift in image restoration from convolutional neural networks (CNNs) to Transformerbased models due to their powerful ability to model long-range pixel interactions. In this paper, we explore the potential of CNNs for image restoration and show that the proposed simple convolutional network architecture, termed ConvIR, can perform on par with or better than the Transformer counterparts. By re-examing the characteristics of 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 our ConvIR delivers state-ofthe- art performance with low computation complexity among 20 benchmark datasets on five representative image restoration tasks, including image dehazing, image motion/defocus deblurring, image deraining, and image desnowing.

Paper PDF Code

Code

c-yn/ConvIR 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 Snow Removal

Results from the paper archive 2025-07-28

Task	Dataset	Model	Metric	Value	Rank at snapshot	Leaderboard	Report
Deblurring	RSBlur	ConvIR	Average PSNR	34.06	#6 of 12	Archive leaderboard	report
Image Deblurring	GoPro	ConvIR	PSNR	33.28	#26 of 55	Archive leaderboard	report
Image Deblurring	GoPro	ConvIR	SSIM	0.963	#26 of 55	Archive leaderboard	report
Image Dehazing	Haze4k	ConvIR	PSNR	34.50	#3 of 11	Archive leaderboard	report
Image Dehazing	Haze4k	ConvIR	SSIM	0.99	#3 of 11	Archive leaderboard	report
Image Dehazing	I-Haze	ConvIR	PSNR	22.44	#3 of 4	Archive leaderboard	report
Image Dehazing	I-Haze	ConvIR	SSIM	0.887	#3 of 4	Archive leaderboard	report
Image Dehazing	O-Haze	ConvIR	PSNR	25.36	#2 of 7	Archive leaderboard	report
Image Dehazing	O-Haze	ConvIR	SSIM	0.780	#2 of 7	Archive leaderboard	report
Image Dehazing	SOTS Indoor	ConvIR	PSNR	42.72	#1 of 34	Archive leaderboard	report
Image Dehazing	SOTS Indoor	ConvIR	SSIM	0.997	#1 of 34	Archive leaderboard	report
Image Dehazing	SOTS Outdoor	ConvIR	PSNR	39.42	#5 of 31	Archive leaderboard	report
Image Dehazing	SOTS Outdoor	ConvIR	SSIM	0.996	#5 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