《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (9): 2975-2983.DOI: 10.11772/j.issn.1001-9081.2024091382
• 多媒体计算与计算机仿真 • 上一篇
景攀峰1, 梁宇栋1,2(), 李超伟3, 郭俊茹1, 郭晋育1
收稿日期:
2024-09-27
修回日期:
2025-01-07
接受日期:
2025-01-13
发布日期:
2025-03-17
出版日期:
2025-09-10
通讯作者:
梁宇栋
作者简介:
景攀峰(1999—),男,山西运城人,硕士研究生,主要研究方向:计算机视觉、图像处理基金资助:
Panfeng JING1, Yudong LIANG1,2(), Chaowei LI3, Junru GUO1, Jinyu GUO1
Received:
2024-09-27
Revised:
2025-01-07
Accepted:
2025-01-13
Online:
2025-03-17
Published:
2025-09-10
Contact:
Yudong LIANG
About author:
JING Panfeng, born in 1999, M. S. candidate. His research interests include computer vision, image processing.Supported by:
摘要:
图像去雾是计算机视觉领域的热点话题之一。由于真实世界中大规模高质量的配对数据集的获取存在成本昂贵、实施困难等问题,现有方法通常利用合成数据对深度学习模型进行全监督训练,这可能会导致模型在真实场景下的泛化性能较差。为了解决真实域和合成域之间的域差异问题,提出一种基于师生学习的半监督图像去雾算法。该算法采用一个半监督的师生学习框架,利用指数移动平均(EMA)策略来更新教师模型,并端到端地进行去雾学习,显著地解决了合成数据与真实数据之间的域偏移问题,并提高了模型在真实有雾场景下的泛化性能。实验结果表明,所提算法在2个合成雾霾图像数据集SOTS(Synthetic Objective Testing Set)、Haze4K和真实雾霾图像数据集URHI(Unannotated Real-world Hazy Images)上取得了较好性能,并获得了更好的去雾视觉效果。
中图分类号:
景攀峰, 梁宇栋, 李超伟, 郭俊茹, 郭晋育. 基于师生学习的半监督图像去雾算法[J]. 计算机应用, 2025, 45(9): 2975-2983.
Panfeng JING, Yudong LIANG, Chaowei LI, Junru GUO, Jinyu GUO. Semi-supervised image dehazing algorithm based on teacher-student learning[J]. Journal of Computer Applications, 2025, 45(9): 2975-2983.
算法 | 年份 | PSNR/dB | SSIM |
---|---|---|---|
DCP[ | 2011 | 18.250 4 | 0.853 2 |
DehazeNet[ | 2016 | 18.144 2 | 0.561 4 |
AOD-Net[ | 2017 | 20.600 3 | 0.858 7 |
GridDehazeNet[ | 2019 | 25.704 1 | 0.931 8 |
FFA-Net[ | 2020 | 27.198 4 | 0.948 2 |
MSBDN[ | 2020 | 27.137 3 | 0.951 7 |
AECR-Net[ | 2021 | 27.904 0 | 0.935 6 |
DehazeFormer[ | 2023 | 30.444 2 | |
MixDehazeNet[ | 2023 | 29.870 8 | 0.961 3 |
FSNet[ | 2024 | 31.271 2 | 0.972 6 |
ConvIR[ | 2024 | 30.823 6 | 0.964 7 |
本文算法 | 2024 | 0.968 0 |
表1 SOTS合成数据集上去雾结果的定量比较
Tab. 1 Quantitative comparison of dehazing results on SOTS synthetic dataset
算法 | 年份 | PSNR/dB | SSIM |
---|---|---|---|
DCP[ | 2011 | 18.250 4 | 0.853 2 |
DehazeNet[ | 2016 | 18.144 2 | 0.561 4 |
AOD-Net[ | 2017 | 20.600 3 | 0.858 7 |
GridDehazeNet[ | 2019 | 25.704 1 | 0.931 8 |
FFA-Net[ | 2020 | 27.198 4 | 0.948 2 |
MSBDN[ | 2020 | 27.137 3 | 0.951 7 |
AECR-Net[ | 2021 | 27.904 0 | 0.935 6 |
DehazeFormer[ | 2023 | 30.444 2 | |
MixDehazeNet[ | 2023 | 29.870 8 | 0.961 3 |
FSNet[ | 2024 | 31.271 2 | 0.972 6 |
ConvIR[ | 2024 | 30.823 6 | 0.964 7 |
本文算法 | 2024 | 0.968 0 |
算法 | 年份 | PSNR/dB | SSIM |
---|---|---|---|
DCP[ | 2011 | 16.926 5 | 0.858 2 |
DehazeNet[ | 2016 | 15.600 5 | 0.534 2 |
AOD-Net[ | 2017 | 17.871 3 | 0.821 9 |
GridDehazeNet[ | 2019 | 21.185 4 | 0.900 1 |
FFA-Net[ | 2020 | 22.109 3 | 0.921 8 |
MSBDN[ | 2020 | 22.932 1 | 0.928 9 |
AECR-Net[ | 2021 | 22.421 5 | 0.927 6 |
DehazeFormer[ | 2023 | 0.946 0 | |
MixDehazeNet[ | 2023 | 23.016 1 | 0.935 3 |
FSNet[ | 2024 | 23.587 0 | |
ConvIR[ | 2024 | 23.236 5 | 0.942 0 |
本文算法 | 2024 | 25.130 6 | 0.951 5 |
表2 Haze4K合成数据集上去雾结果的定量比较
Tab. 2 Quantitative comparison of dehazing results on Haze4K synthetic dataset
算法 | 年份 | PSNR/dB | SSIM |
---|---|---|---|
DCP[ | 2011 | 16.926 5 | 0.858 2 |
DehazeNet[ | 2016 | 15.600 5 | 0.534 2 |
AOD-Net[ | 2017 | 17.871 3 | 0.821 9 |
GridDehazeNet[ | 2019 | 21.185 4 | 0.900 1 |
FFA-Net[ | 2020 | 22.109 3 | 0.921 8 |
MSBDN[ | 2020 | 22.932 1 | 0.928 9 |
AECR-Net[ | 2021 | 22.421 5 | 0.927 6 |
DehazeFormer[ | 2023 | 0.946 0 | |
MixDehazeNet[ | 2023 | 23.016 1 | 0.935 3 |
FSNet[ | 2024 | 23.587 0 | |
ConvIR[ | 2024 | 23.236 5 | 0.942 0 |
本文算法 | 2024 | 25.130 6 | 0.951 5 |
模块 | PSNR/dB | SSIM |
---|---|---|
Baseline | 29.789 1 | 0.965 7 |
Baseline+PCSAB | 30.609 9 | 0.966 2 |
Baseline+师生学习 | 30.259 2 | 0.963 9 |
Baseline+PCSAB+师生学习 | 31.032 1 | 0.968 0 |
表3 PCSAB与师生学习对模型性能的影响
Tab. 3 Influence of PCSAB and teacher-student learning on model performance
模块 | PSNR/dB | SSIM |
---|---|---|
Baseline | 29.789 1 | 0.965 7 |
Baseline+PCSAB | 30.609 9 | 0.966 2 |
Baseline+师生学习 | 30.259 2 | 0.963 9 |
Baseline+PCSAB+师生学习 | 31.032 1 | 0.968 0 |
[1] | XIONG Y, LI Z, CHEN Y, et al. Efficient deformable ConvNets: rethinking dynamic and sparse operator for vision applications [C]// Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2024: 5652-5661. |
[2] | XIAO D, MENG Q, LI S, et al. Improving Transformers with dynamically composable multi-head attention [C]// Proceedings of the 41st International Conference on Machine Learning. New York: JMLR.org, 2024: 54300-54318. |
[3] | KIM T K, PAIK J K, KANG B S. Contrast enhancement system using spatially adaptive histogram equalization with temporal filtering [J]. IEEE Transactions on Consumer Electronics, 1998, 44(1): 82-87. |
[4] | FAN Y, ZHANG L, GUO H, et al. Image processing for laser imaging using adaptive homomorphic filtering and total variation[J]. Photonics, 2020, 7(2): No.30. |
[5] | DAUBECHIES I. The wavelet transform, time-frequency localization and signal analysis [J]. IEEE Transactions on Information Theory, 1990, 36(5): 961-1005. |
[6] | RAHMAN Z U, JOBSON D J, WOODELL G A. Retinex processing for automatic image enhancement [J]. Journal of Electronic Imaging, 2004, 13(1): No.1636183. |
[7] | NARASIMHAN S G, NAYAR S K. Chromatic framework for vision in bad weather [C]// Proceedings of the 2000 IEEE Conference on Computer Vision and Pattern Recognition — Volume 1. Piscataway: IEEE, 2000: 598-605. |
[8] | NARASIMHAN S G, NAYAR S K. Vision and the atmosphere [J]. International Journal of Computer Vision, 2002, 48(3): 233-254. |
[9] | TAN R T. Visibility in bad weather from a single image [C]// Proceedings of the 2008 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2008: 1-8. |
[10] | FATTAL R. Single image dehazing [J]. ACM Transactions on Graphics, 2008, 27(3): No.72. |
[11] | HE K, SUN J, TANG X. Single image haze removal using dark channel prior [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341-2353. |
[12] | CAI B, XU X, JIA K, et al. DehazeNet: an end-to-end system for single image haze removal [J]. IEEE Transactions on Image Processing, 2016, 25(11): 5187-5198. |
[13] | LI B, PENG X, WANG Z, et al. An all-in-one network for dehazing and beyond [EB/OL]. [2024-01-14]. . |
[14] | LIU X, MA Y, SHI Z, et al. GridDehazeNet: attention-based multi-scale network for image dehazing [C]// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 7313-7322. |
[15] | QIN X, WANG Z, BAI Y, et al. FFA-Net: feature fusion attention network for single image dehazing [C]// Proceedings of the 34th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2020: 11908-11915. |
[16] | DONG H, PAN J, XIANG L, et al. Multi-scale boosted dehazing network with dense feature fusion [C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 2154-2164. |
[17] | WU H, QU Y, LIN S, et al. Contrastive learning for compact single image dehazing [C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 10546-10555. |
[18] | LIU Y, ZHU L, PEI S, et al. From synthetic to real: image dehazing collaborating with unlabeled real data [C]// Proceedings of the 29th ACM International Conference on Multimedia. New York: ACM, 2021: 50-58. |
[19] | SONG Y, HE Z, QIAN H, et al. Vision Transformers for single image dehazing [J]. IEEE Transactions on Image Processing, 2023, 32: 1927-1941. |
[20] | LU L, XIONG Q, XU B, et al. MixDehazeNet: mix structure block for image dehazing network [C]// Proceedings of the 2024 International Joint Conference on Neural Networks. Piscataway: IEEE, 2024: 1-10. |
[21] | CUI Y, REN W, CAO X, et al. Image restoration via frequency selection [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(2): 1093-1108. |
[22] | CUI Y, REN W, CAO X, et al. Revitalizing convolutional network for image restoration [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(12): 9423-9438. |
[23] | YANG W, WANG S, FANG Y, et al. From fidelity to perceptual quality: a semi-supervised approach for low-light image enhancement [C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 3060-3069. |
[24] | YASARLA R, SINDAGI V A, PATEL V M. Syn2Real transfer learning for image deraining using Gaussian processes [C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 2723-2733. |
[25] | SHAO Y, LI L, REN W, et al. Domain adaptation for image dehazing [C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 2805-2814. |
[26] | YUE Z, XIE J, ZHAO Q, et al. Semi-supervised video deraining with dynamical rain generator [C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 642-652. |
[27] | HUANG H, YU A, HE R. Memory oriented transfer learning for semi-supervised image deraining [C]// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 7728-7737. |
[28] | WANG J, WANG F, YIN D. Feature decoupled autoencoder: semi-supervised learning for image dehazing [C]// Proceedings of the 2022 IEEE International Conference on Multimedia and Expo. Piscataway: IEEE, 2022: 1-6. |
[29] | SUN Z, ZHANG Y, BAO F, et al. SADnet: semi-supervised single image dehazing method based on an attention mechanism[J]. ACM Transactions on Multimedia Computing, Communications, and Applications, 2022, 18(2): No.58. |
[30] | YASARLA R, PRIEBE C E, PATEL V M. ART-SS: an adaptive rejection technique for semi-supervised restoration for adverse weather-affected images [C]// Proceedings of the 2022 European Conference on Computer Vision, LNCS 13678. Cham: Springer, 2022: 699-718. |
[31] | CUI X, WANG C, REN D, et al. Semi-supervised image deraining using knowledge distillation [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2022, 32(12): 8327-8341. |
[32] | DONG Y, LI Y, DONG Q, et al. Semi-supervised domain alignment learning for single image dehazing [J]. IEEE Transactions on Cybernetics, 2023, 53(11): 7238-7250. |
[33] | LIU Y, YAN Z, CHEN S, et al. NightHazeFormer: single nighttime haze removal using prior query Transformer [C]// Proceedings of the 31st ACM International Conference on Multimedia. New York: ACM, 2023: 4119-4128. |
[34] | LI B, REN W, FU D, et al. Benchmarking single-image dehazing and beyond [J]. IEEE Transactions on Image Processing, 2019, 28(1): 492-505. |
[35] | KINGMA D P, BA J L. Adam: a method for stochastic optimization [EB/OL]. [2024-07-05].. |
[1] | 张宏俊, 潘高军, 叶昊, 陆玉彬, 缪宜恒. 结合深度学习和张量分解的多源异构数据分析方法[J]. 《计算机应用》唯一官方网站, 2025, 45(9): 2838-2847. |
[2] | 李进, 刘立群. 基于残差Swin Transformer的SAR与可见光图像融合[J]. 《计算机应用》唯一官方网站, 2025, 45(9): 2949-2956. |
[3] | 殷兵, 凌震华, 林垠, 奚昌凤, 刘颖. 兼容缺失模态推理的情感识别方法[J]. 《计算机应用》唯一官方网站, 2025, 45(9): 2764-2772. |
[4] | 葛丽娜, 王明禹, 田蕾. 联邦学习的高效性研究综述[J]. 《计算机应用》唯一官方网站, 2025, 45(8): 2387-2398. |
[5] | 廖炎华, 鄢元霞, 潘文林. 基于YOLOv9的交通路口图像的多目标检测算法[J]. 《计算机应用》唯一官方网站, 2025, 45(8): 2555-2565. |
[6] | 彭鹏, 蔡子婷, 刘雯玲, 陈才华, 曾维, 黄宝来. 基于CNN和双向GRU混合孪生网络的语音情感识别方法[J]. 《计算机应用》唯一官方网站, 2025, 45(8): 2515-2521. |
[7] | 张硕, 孙国凯, 庄园, 冯小雨, 王敬之. 面向区块链节点分析的eclipse攻击动态检测方法[J]. 《计算机应用》唯一官方网站, 2025, 45(8): 2428-2436. |
[8] | 索晋贤, 张丽萍, 闫盛, 王东奇, 张雅雯. 可解释的深度知识追踪方法综述[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2043-2055. |
[9] | 王震洲, 郭方方, 宿景芳, 苏鹤, 王建超. 面向智能巡检的视觉模型鲁棒性优化方法[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2361-2368. |
[10] | 齐巧玲, 王啸啸, 张茜茜, 汪鹏, 董永峰. 基于元学习的标签噪声自适应学习算法[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2113-2122. |
[11] | 赵小阳, 许新征, 李仲年. 物联网应用中的可解释人工智能研究综述[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2169-2179. |
[12] | 李岚皓, 严皓钧, 周号益, 孙庆赟, 李建欣. 基于神经网络的多尺度信息融合时间序列长期预测模型[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 1776-1783. |
[13] | 花天辰, 马晓宁, 智慧. 基于浅层人工神经网络的可移植执行恶意软件静态检测模型[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 1911-1921. |
[14] | 牛四杰, 刘昱良. 基于知识蒸馏双分支结构的视网膜病变辅助诊断方法[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1410-1414. |
[15] | 王文鹏, 秦寅畅, 师文轩. 工业缺陷检测无监督深度学习方法综述[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1658-1670. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||