Papers › Reversible Decoupling Network for Single Image Reflection Removal

Reversible Decoupling Network for Single Image Reflection Removal

10 Oct 2024CVPR 2025 1arXiv:2410.08063archive 2025-07-28

Hao Zhao, Mingjia Li, Qiming Hu, Xiaojie Guo

Recent deep-learning-based approaches to single-image reflection removal have shown promising advances, primarily for two reasons: 1) the utilization of recognition-pretrained features as inputs, and 2) the design of dual-stream interaction networks. However, according to the Information Bottleneck principle, high-level semantic clues tend to be compressed or discarded during layer-by-layer propagation. Additionally, interactions in dual-stream networks follow a fixed pattern across different layers, limiting overall performance. To address these limitations, we propose a novel architecture called Reversible Decoupling Network (RDNet), which employs a reversible encoder to secure valuable information while flexibly decoupling transmission- and reflection-relevant features during the forward pass. Furthermore, we customize a transmission-rate-aware prompt generator to dynamically calibrate features, further boosting performance. Extensive experiments demonstrate the superiority of RDNet over existing SOTA methods on five widely-adopted benchmark datasets. RDNet achieves the best performance in the NTIRE 2025 Single Image Reflection Removal in the Wild Challenge in both fidelity and perceptual comparison. Our code is available at https://github.com/lime-j/RDNet

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

lime-j/RDNet officialmentioned in papermentioned on GitHubpytorch 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

Reflection Removal

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Reflection Removal Nature RDNet PSNR 26.21 #2 of 5 Archive leaderboard report
Reflection Removal Nature RDNet SSIM 0.842 #2 of 5 Archive leaderboard report
Reflection Removal Nature Zhu et al. PSNR 26.04 #3 of 5 Archive leaderboard report
Reflection Removal Nature Zhu et al. SSIM 0.846 #3 of 5 Archive leaderboard report
Reflection Removal Real20 RDNet PSNR 25.58 #1 of 8 Archive leaderboard report
Reflection Removal Real20 RDNet SSIM 0.846 #1 of 8 Archive leaderboard report
Reflection Removal SIR^2(Objects) RDNet PSNR 26.78 #1 of 7 Archive leaderboard report
Reflection Removal SIR^2(Objects) RDNet SSIM 0.921 #1 of 7 Archive leaderboard report
Reflection Removal SIR^2(Objects) Zhu et al. SSIM 0.931 #7 of 7 Archive leaderboard report
Reflection Removal SIR^2(Postcard) RDNet PSNR 26.33 #1 of 6 Archive leaderboard report
Reflection Removal SIR^2(Postcard) RDNet SSIM 0.922 #1 of 6 Archive leaderboard report
Reflection Removal SIR^2(Wild) RDNet PSNR 27.7 #1 of 6 Archive leaderboard report
Reflection Removal SIR^2(Wild) RDNet SSIM 0.915 #1 of 6 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

RDNet

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