AFE‐Dehaze: Image Dehazing Method Based on Adaptive Feature Enhancement Contrastive Learning

Lanqing Zhang; Zhigao Cui; Yanzhao Su; Nian Wang; Yunwei Lan; Liangyu Zhu; Cheng Chen

doi:10.1049/ipr2.70230

journal article Open Access Jan 01, 2025

AFE‐Dehaze: Image Dehazing Method Based on Adaptive Feature Enhancement Contrastive Learning

Lanqing Zhang

Zhigao Cui Yanzhao Su Nian Wang

Yunwei Lan

Liangyu Zhu

Cheng Chen

IET Image Processing Vol. 19 No. 1 · Institution of Engineering and Technology (IET)

View at Publisher Save 10.1049/ipr2.70230

Abstract

ABSTRACT
To address the issue of recovery imbalance caused by the spatial heterogeneity of haze concentration in real scenarios, this paper proposes an adaptive feature enhanced contrastive learning framework (AFE‐Dehaze). The framework achieves breakthroughs through three major collaborative mechanisms: (1) a hierarchical multi‐scale fusion architecture that combines diffusion convolution and channel attention, preserving edge textures (such as leaf veins and building contours) in thin haze areas, while semantically guiding the reconstruction of structural details in dense haze areas, improving texture retention by 18% compared to traditional U‐Net; (2) a concentration‐sensitive contrastive learning paradigm that uses pre‐trained VGG features as semantic anchors, applying pixel‐level constraints in thin haze and feature space constraints in dense haze, which reduces color distortion () by 23%, significantly outperforming methods like refusion; (3) a gradient dynamic balancing strategy that automatically adjusts the optimization direction by analyzing positive and negative sample gradient contributions, enhancing PSNR by 1.2dB and SSIM by 0.05 in non‐uniform haze scenarios. Experiments on a mixed dataset (RESIDE OTS real scenes) demonstrate that AFE‐Dehaze achieves an average PSNR of 28.7dB and SSIM of 0.91, especially improving structural similarity in dense haze areas by 9% compared to Mamba, validating its generalization capability in complex haze environments. This framework provides a solution that balances accuracy and robustness for dehazing in real scenarios such as vehicular vision and remote sensing imaging.

Topics

No keywords indexed for this article. Browse by subject →

References

52

[1]

10.1109/tpami.2024.3416731

[2]

10.1109/tpami.2021.3070580

[3]

10.1109/tip.2024.3362153

[4]

10.1109/tcsvt.2024.3414677

[5]

10.1109/tmm.2021.3110483

[6]

Image Restoration via Frequency Selection

Yuning Cui, Wenqi Ren, Xiaochun Cao et al.

IEEE Transactions on Pattern Analysis and Machine... 10.1109/tpami.2023.3330416

[7]

Y.KimandC.Yim “Image Dehaze Method Using Depth Map Estimation Network Based on Atmospheric Scattering Model ” in2020 International Conference on Electronics Information and Communication (ICEIC)(IEEE 2020) 1–3. 10.1109/iceic49074.2020.9051031

[8]

Single Image Haze Removal Using Dark Channel Prior

Kaiming He, Jian Sun, Xiaoou Tang

IEEE Transactions on Pattern Analysis and Machine... 10.1109/tpami.2010.168

[9]

10.1109/tip.2015.2446191

[10]

B.Li X.Peng Z.Wang J.Xu andD.Feng “AOD‐Net: All‐in‐One Dehazing Network ” in2017 IEEE International Conference on Computer Vision (ICCV)(IEEE 2017) 4780–4788. 10.1109/iccv.2017.511

[11]

10.1109/tip.2016.2598681

[12]

10.1109/tits.2022.3225797

[13]

Y.Zhang S.Zhou andH.Li “Depth Information Assisted Collaborative Mutual Promotion Network for Single Image Dehazing ” in2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)(IEEE 2024) 2846–2855. 10.1109/cvpr52733.2024.00275

[14]

10.1109/tip.2023.3245994

[15]

10.1109/tgrs.2022.3204890

[16]

Z.Wang H.Zhao J.Peng L.Yao andK.Zhao “ODCR: Orthogonal Decoupling Contrastive Regularization for Unpaired Image Dehazing ” in2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)(2024) 25479–25489. 10.1109/cvpr52733.2024.02407

[17]

GUDCP: Generalization of Underwater Dark Channel Prior for Underwater Image Restoration

Zheng Liang, Xueyan Ding, Yafei Wang et al.

IEEE Transactions on Circuits and Systems for Vide... 10.1109/tcsvt.2021.3114230

[18]

10.1109/tits.2020.3030673

[19]

D.Berman T.Treibitz andS.Avidan “Non‐Local Image Dehazing ” in2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)(IEEE 2016) 1674–1682. 10.1109/cvpr.2016.185

[20]

10.1109/tmm.2022.3181447

[21]

D.Berman S.Avidan et al. “Non‐Local Image Dehazing ” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition(IEEE 2016) 1674–1682. 10.1109/cvpr.2016.185

[22]

10.1145/2651362

[23]

He K. "Single Image Haze Removal Using Dark Channel Prior" IEEE Transactions on Pattern Analysis and Machine Intelligence (2010)

[24]

Q.Zhu J.Mai andL.Shao “Single Image Dehazing Using Color Attenuation prior ” inBMVC Vol.4(Springer 2014) 1674–1682.

[25]

10.1109/tip.2016.2598681

[26]

10.1007/s11263-019-01235-8

[27]

X.Liu Y.Ma Z.Shi andJ.Chen “Griddehazenet: Attention‐Based Multi‐Scale Network for Image Dehazing ” in2019 IEEE/CVF International Conference on Computer Vision (ICCV)(IEEE 2019) 7313–7322. 10.1109/iccv.2019.00741

[28]

10.1109/tip.2024.3354108

[29]

X.Qin Z.Wang Y.Bai X.Xie andH.Jia “Ffa‐net: Feature Fusion Attention Network for Single Image Dehazing ” CoRR abs/1911.07559 (2019).

[30]

H.Dong J.Pan L.Xiang et al. “Multi‐Scale Boosted Dehazing Network With Dense Feature Fusion ” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2020) 2157–2167. 10.1109/cvpr42600.2020.00223

[31]

10.1007/s11760-022-02252-w

[32]

10.1109/lsp.2025.3551201

[33]

10.1109/tip.2019.2952032

[34]

10.1109/tip.2021.3060873

[35]

Y.Shao L.Li W.Ren C.Gao andN.Sang “Domain Adaptation for Image Dehazing ” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2020) 2808–2817. 10.1109/cvpr42600.2020.00288

[36]

Y.Yang C.Wang R.Liu L.Zhang X.Guo andD.Tao “Self‐Augmented Unpaired Image Dehazing via Density and Depth Decomposition ” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 2037–2046. 10.1109/cvpr52688.2022.00208

[37]

C.‐L.Guo Q.Yan S.Anwar R.Cong W.Ren andC.Li “Image Dehazing Transformer with Transmission‐Aware 3D Position Embedding ” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2022) 5812–5820. 10.1109/cvpr52688.2022.00572

[38]

Z.Chen Y.Wang Y.Yang andD.Liu “Psd: Principled Synthetic‐to‐Real Dehazing Guided by Physical Priors ” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2021) 7180–7189. 10.1109/cvpr46437.2021.00710

[39]

10.1109/tip.2019.2952690

[40]

10.1109/tits.2022.3170328

[41]

10.1016/j.engappai.2024.109012

[42]

Y.Lan Z.Cui C.Liu et al. “Exploiting Diffusion Prior for Real‐World Image Dehazing with Unpaired Training ” inProceedings of the AAAI Conference on Artificial Intelligence39 no.4(2025):4455–4463. 10.1609/aaai.v39i4.32469

[43]

10.1109/tmm.2025.3543063

[44]

10.1016/j.patcog.2023.109700

[45]

X.Cong J.Gui J.Zhang J.Hou andH.Shen “A Semi‐Supervised Nighttime Dehazing Baseline with Spatial‐Frequency Aware and Realistic Brightness Constraint ” in2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)(IEEE 2024) 2631–2640. 10.1109/cvpr52733.2024.00254

[46]

X.Song J.Dai Z.Fu andD.Tan “Optimization Improvement of AECR‐Net Image Dehazing Network ” in2024 IEEE 6th International Conference on Power Intelligent Computing and Systems (ICPICS)(IEEE 2024) 1428–1436. 10.1109/icpics62053.2024.10796615

[47]

10.1007/s00530‐024‐01462‐1

[48]

10.1007/s00371‐024‐03511‐2

[49]

R.‐Q.Wu Z.‐P.Duan C.‐L.Guo Z.Chai andC.Li “RIDCP: Revitalizing Real Image Dehazing via High‐Quality Codebook Priors ” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(IEEE 2023) 22282–22291. 10.1109/cvpr52729.2023.02134

[50]

U-Net: Convolutional Networks for Biomedical Image Segmentation

Olaf Ronneberger, Philipp Fischer, Thomas Brox

Lecture Notes in Computer Science 10.1007/978-3-319-24574-4_28

Showing 50 of 52 references

Metrics

1

Citations

52

References

Details

Published: Jan 01, 2025
Vol/Issue: 19(1)
License: View

Authors

L

Lanqing Zhang