《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (3): 922-930.DOI: 10.11772/j.issn.1001-9081.2023030367
所属专题: 多媒体计算与计算机仿真
收稿日期:
2023-04-07
修回日期:
2023-06-08
接受日期:
2023-06-13
发布日期:
2023-09-07
出版日期:
2024-03-10
通讯作者:
刘威
作者简介:
江锐(1998—),男,湖北黄冈人,硕士研究生,主要研究方向:图像处理、深度学习基金资助:
Rui JIANG1,2, Wei LIU1,2(), Cheng CHEN1,2, Tao LU1,2
Received:
2023-04-07
Revised:
2023-06-08
Accepted:
2023-06-13
Online:
2023-09-07
Published:
2024-03-10
Contact:
Wei LIU
About author:
JIANG Rui, born in 1998, M. S. candidate. His research interests include image processing, deep learning.Supported by:
摘要:
现有的基于学习的单幅图像去雨网络大都关注雨天图像中雨痕对于视觉成像的影响,而忽略了雨天环境下由于空气中湿度的增加所产生的雾气对视觉成像的影响,因此造成去雨后图像的生成质量低、纹理细节信息模糊等问题。针对该问题,提出一种非对称端到端的无监督图像去雨网络模型,该模型主要包含雨雾去除网络、雨雾特征提取网络和雨雾生成网络,并由它们组成两个不同数据域映射转换模块:Rain-Clean-Rain和Clean-Rain-Clean。上述三个子网络构成并行的两条转换路径:去雨路径和雨雾特征提取路径。在雨雾特征提取路径上,提出一种基于全局和局部注意力机制的雨雾感知提取网络,利用雨雾特征存在的全局自相似性和局部差异性学习雨-雾相关特征;在去雨路径上,引入雨天图像退化模型和上述提取的雨雾相关特征作为先验知识以增强雨雾图像生成的能力,从而约束雨雾去除网络,提高它从雨天数据域到无雨数据域的映射转换能力。在不同雨天图像数据集上的实验结果表明,与较先进的去雨方法CycleDerain相比,在合成雨雾数据集HeavyRain上所提方法的峰值信噪比(PSNR)提升了31.55%,能适应不同的雨天场景,具有更好的泛化性,并且能更好地复原图像的细节和纹理信息。
中图分类号:
江锐, 刘威, 陈成, 卢涛. 非对称端到端的无监督图像去雨网络[J]. 计算机应用, 2024, 44(3): 922-930.
Rui JIANG, Wei LIU, Cheng CHEN, Tao LU. Asymmetric unsupervised end-to-end image deraining network[J]. Journal of Computer Applications, 2024, 44(3): 922-930.
模型 | HeavyRain | Rain800 | Rain1200 | |||||||
---|---|---|---|---|---|---|---|---|---|---|
PSNR/dB | SSIM | LPIPS | PSNR/dB | SSIM | LPIPS | PSNR/dB | SSIM | LPIPS | ||
有监督去雨模型 | Detail | 16.40 | 0.501 | 0.709 | 22.00 | 0.780 | 0.131 | 21.53 | 0.798 | 0.091 |
LPNet | 15.88 | 0.512 | 0.602 | 20.92 | 0.756 | 0.188 | 22.22 | 0.781 | 0.127 | |
SPANet | 16.65 | 0.506 | 0.572 | 22.68 | 0.787 | 0.205 | 23.53 | 0.770 | 0.159 | |
DualGCN | 15.31 | 0.655 | 0.420 | 22.22 | 0.782 | 0.107 | 27.34 | 0.863 | 0.076 | |
半监督去雨模型 | SEMI | 16.45 | 0.398 | 0.607 | 21.16 | 0.731 | 0.138 | 22.50 | 0.724 | 0.185 |
无监督去雨模型 | CycleDerain | 18.64 | 0.532 | 0.359 | 21.75 | 0.721 | 0.210 | 21.30 | 0.694 | 0.308 |
本文方法 | 24.52 | 0.821 | 0.062 | 23.19 | 0.789 | 0.104 | 25.73 | 0.849 | 0.070 |
表1 不同方法在合成雨天数据集上的去雨效果的客观评价指标
Tab. 1 Objective evaluation indexes of deraining effects by different methods on synthetic rain datasets
模型 | HeavyRain | Rain800 | Rain1200 | |||||||
---|---|---|---|---|---|---|---|---|---|---|
PSNR/dB | SSIM | LPIPS | PSNR/dB | SSIM | LPIPS | PSNR/dB | SSIM | LPIPS | ||
有监督去雨模型 | Detail | 16.40 | 0.501 | 0.709 | 22.00 | 0.780 | 0.131 | 21.53 | 0.798 | 0.091 |
LPNet | 15.88 | 0.512 | 0.602 | 20.92 | 0.756 | 0.188 | 22.22 | 0.781 | 0.127 | |
SPANet | 16.65 | 0.506 | 0.572 | 22.68 | 0.787 | 0.205 | 23.53 | 0.770 | 0.159 | |
DualGCN | 15.31 | 0.655 | 0.420 | 22.22 | 0.782 | 0.107 | 27.34 | 0.863 | 0.076 | |
半监督去雨模型 | SEMI | 16.45 | 0.398 | 0.607 | 21.16 | 0.731 | 0.138 | 22.50 | 0.724 | 0.185 |
无监督去雨模型 | CycleDerain | 18.64 | 0.532 | 0.359 | 21.75 | 0.721 | 0.210 | 21.30 | 0.694 | 0.308 |
本文方法 | 24.52 | 0.821 | 0.062 | 23.19 | 0.789 | 0.104 | 25.73 | 0.849 | 0.070 |
方法 | Real147 | RID | ||
---|---|---|---|---|
NIQE | SSEQ | NIQE | SSEQ | |
Detail | 4.276 | 28.607 | 4.326 | 27.535 |
LPNet | 4.831 | 31.509 | 4.797 | 31.265 |
SPANet | 3.987 | 30.214 | 3.591 | 27.969 |
DualGCN | 3.801 | 29.801 | 3.486 | 27.126 |
SEMI | 3.751 | 30.423 | 3.524 | 30.506 |
CycleDerain | 4.361 | 34.146 | 4.560 | 38.009 |
本文方法 | 3.723 | 28.481 | 3.432 | 28.949 |
表2 不同方法在真实雨天图像上去雨效果的客观评价指标
Tab. 2 Objective evaluation indexes of deraining effects ofdifferent methods on real rain images
方法 | Real147 | RID | ||
---|---|---|---|---|
NIQE | SSEQ | NIQE | SSEQ | |
Detail | 4.276 | 28.607 | 4.326 | 27.535 |
LPNet | 4.831 | 31.509 | 4.797 | 31.265 |
SPANet | 3.987 | 30.214 | 3.591 | 27.969 |
DualGCN | 3.801 | 29.801 | 3.486 | 27.126 |
SEMI | 3.751 | 30.423 | 3.524 | 30.506 |
CycleDerain | 4.361 | 34.146 | 4.560 | 38.009 |
本文方法 | 3.723 | 28.481 | 3.432 | 28.949 |
模型 | 不同图像的NIQE | |
---|---|---|
图像1 | 图像2 | |
模型A | 3.501 | 4.502 |
模型B | 3.763 | 4.412 |
模型C | 3.932 | 4.425 |
模型D | 3.279 | 4.031 |
本文模型 | 3.268 | 4.016 |
表3 消融实验的客观评价指标
Tab. 3 Objective evaluation indexes of ablation experiments
模型 | 不同图像的NIQE | |
---|---|---|
图像1 | 图像2 | |
模型A | 3.501 | 4.502 |
模型B | 3.763 | 4.412 |
模型C | 3.932 | 4.425 |
模型D | 3.279 | 4.031 |
本文模型 | 3.268 | 4.016 |
方法 | 图像1 | 图像2 | 图像3 | 图像4 | ||||
---|---|---|---|---|---|---|---|---|
NIQE | SSEQ | NIQE | SSEQ | NIQE | SSEQ | NIQE | SSEQ | |
DCPDN | 2.387 | 19.706 | 3.828 | 10.234 | 5.013 | 24.140 | 3.968 | 21.477 |
DehazeNet | 2.897 | 8.608 | 3.247 | 22.409 | 4.832 | 16.333 | 3.790 | 50.192 |
GCANet | 2.456 | 11.689 | 3.619 | 24.886 | 4.456 | 26.376 | 3.323 | 33.272 |
IDE | 2.203 | 4.844 | 3.003 | 14.056 | 3.347 | 27.881 | 3.661 | 52.372 |
本文方法 | 2.143 | 2.677 | 2.788 | 7.901 | 3.081 | 16.222 | 2.982 | 17.940 |
表4 不同方法对图9中雾图去雾结果的客观评价
Tab. 4 Objective evaluation of defogging results for Fig. 9 by different methods
方法 | 图像1 | 图像2 | 图像3 | 图像4 | ||||
---|---|---|---|---|---|---|---|---|
NIQE | SSEQ | NIQE | SSEQ | NIQE | SSEQ | NIQE | SSEQ | |
DCPDN | 2.387 | 19.706 | 3.828 | 10.234 | 5.013 | 24.140 | 3.968 | 21.477 |
DehazeNet | 2.897 | 8.608 | 3.247 | 22.409 | 4.832 | 16.333 | 3.790 | 50.192 |
GCANet | 2.456 | 11.689 | 3.619 | 24.886 | 4.456 | 26.376 | 3.323 | 33.272 |
IDE | 2.203 | 4.844 | 3.003 | 14.056 | 3.347 | 27.881 | 3.661 | 52.372 |
本文方法 | 2.143 | 2.677 | 2.788 | 7.901 | 3.081 | 16.222 | 2.982 | 17.940 |
1 | 李超, 黄新宇, 王凯. 基于特征融合和自学习锚框的高分辨率图像小目标检测算法[J]. 电子学报, 2022, 50(7): 1684-1695. 10.12263/DZXB.20200917 |
LI C, HUANG X Y, WANG K. Small target detection of high-resolution images based on feature fusion and learnable anchor [J]. Acta Electronica Sinica, 2022, 50(7): 1684-1695. 10.12263/DZXB.20200917 | |
2 | 任莎莎, 刘琼. 小目标特征增强图像分割算法[J]. 电子学报, 2022, 50(8):1894-1904. 10.12263/DZXB.20211123 |
REN S S, LIU Q. A tiny target feature enhancement algorithm for image segmentation [J]. Acta Electronica Sinica, 2022, 50(8): 1894-1904. 10.12263/DZXB.20211123 | |
3 | KANG L-W, LIN C-W, FU Y-H. Automatic single-image-based rain streaks removal via image decomposition [J]. IEEE Transactions on Image Processing, 2012, 21(4): 1742-1755. 10.1109/tip.2011.2179057 |
4 | LUO Y, XU Y, JI H. Removing rain from a single image via discriminative sparse coding [C]// Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2015: 3397-3405. 10.1109/iccv.2015.388 |
5 | ZHU L, FU C-W, LISCHINSKI D, et al. Joint bi-layer optimization for single-image rain streak removal [C]// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 2545-2553. 10.1109/iccv.2017.276 |
6 | LI Y, TAN R T, GUO X, et al. Rain streak removal using layer priors[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 2736-2744. 10.1109/cvpr.2016.299 |
7 | FU X, HUANG J, DING X, et al. Clearing the skies: a deep network architecture for single-image rain removal [J]. IEEE Transactions on Image Processing, 2017, 26(6): 2944-2956. 10.1109/tip.2017.2691802 |
8 | REN D, ZUO W, HU Q, et al. Progressive image deraining networks: a better and simpler baseline [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 3937-3946. 10.1109/cvpr.2019.00406 |
9 | ZHANG H, SINDAGI V, PATEL V M. Image de-raining using a conditional generative adversarial network [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2019, 30(11): 3943-3956. 10.1109/tcsvt.2019.2920407 |
10 | YANG W, TAN R T, FENG J, et al. Deep joint rain detection and removal from a single image [C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 1357-1366. 10.1109/cvpr.2017.183 |
11 | LI R, L-F CHEONG, TAN R T. Heavy rain image restoration: integrating physics model and conditional adversarial learning [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 1633-1642. 10.1109/cvpr.2019.00173 |
12 | 盖杉, 王俊生. 基于深度学习的非局部注意力增强网络图像去雨算法研究[J]. 电子学报, 2020, 48(10): 1899-1908. 10.3969/j.issn.0372-2112.2020.10.004 |
GAI S, WANG J S. Image raindrop algorithm research using nonlocal attention enhanced network based on deep learning [J]. Acta Electronica Sinica, 2020, 48(10): 1899-1908. 10.3969/j.issn.0372-2112.2020.10.004 | |
13 | GOODFELLOW I, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets [C]// Proceedings of the 27th International Conference on Neural Information Processing Systems. Cambridge: MIT Press, 2014: 2672-2680. |
14 | ZHU H, PENG X, ZHOU J T, et al. Singe image rain removal with unpaired information: A differentiable programming perspective[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 9332-9339. 10.1609/aaai.v33i01.33019332 |
15 | JIN X, CHEN Z, LIN J, et al. Unsupervised single image deraining with self-supervised constraints [C]// Proceedings of the 2019 IEEE International Conference on Image Processing. Piscataway: IEEE, 2019: 2761-2765. 10.1109/icip.2019.8803238 |
16 | WEI Y, ZHANG Z, WANG Y, et al. DerainCycleGAN: rain attentive CycleGAN for single image deraining and rainmaking [J]. IEEE Transactions on Image Processing, 2021, 30: 4788-4801. 10.1109/tip.2021.3074804 |
17 | ZHU J-Y, PARK T, ISOLA P, et al. Unpaired image-to-image translation using cycle-consistent adversarial networks [C]// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 2223-2232. 10.1109/iccv.2017.244 |
18 | HU J, SHEN L, SUN G. Squeeze-and-excitation networks [C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 7132-7141. 10.1109/cvpr.2018.00745 |
19 | WOO S, PARK J, LEE J-Y, et al. CBAM: convolutional block attention module [C]// Proceedings of the 2018 European Conference on Computer Vision. Cham: Springer, 2018: 3-19. 10.1007/978-3-030-01234-2_1 |
20 | WANG T, YANG X, XU K, et al. Spatial attentive single-image deraining with a high quality real rain dataset [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 12270-12279. 10.1109/cvpr.2019.01255 |
21 | WANG C, WU Y, SU Z, et al. Joint self-attention and scale-aggregation for self-calibrated deraining network [C]// Proceedings of the 28th ACM International Conference on Multimedia. New York: ACM, 2020: 2517-2525. 10.1145/3394171.3413559 |
22 | ZHANG H, PATEL V M. Density-aware single image de-raining using a multi-stream dense network [C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 695-704. 10.1109/cvpr.2018.00079 |
23 | WEI W, MENG D, ZHAO Q, et al. Semi-supervised transfer learning for image rain removal [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 3877-3886. 10.1109/cvpr.2019.00400 |
24 | LI S, ARAUJO I B, REN W, et al. Single image deraining: a comprehensive benchmark analysis [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 3838-3847. 10.1109/cvpr.2019.00396 |
25 | ZHANG Y, DING L, SHARMA G. HazeRD: an outdoor scene dataset and benchmark for single image dehazing [C]// Proceedings of the 2017 IEEE International Conference on Image Processing. Piscataway: IEEE, 2017: 3205-3209. 10.1109/icip.2017.8296874 |
26 | FU X, HUANG J, ZENG D, et al. Removing rain from single images via a deep detail network [C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 3855-3863. 10.1109/cvpr.2017.186 |
27 | FU X, LIANG B, HUANG Y, et al. Lightweight pyramid networks for image deraining [J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 31(6): 1794-1807. 10.1109/tnnls.2019.2926481 |
28 | FU X, QI Q, ZHA Z-J, et al. Rain streak removal via dual graph convolutional network [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(2): 1352-1360. 10.1609/aaai.v35i2.16224 |
29 | HUYNH-THU Q, GHANBARI M. Scope of validity of PSNR in image/video quality assessment [J]. Electronics Letters, 2008, 44(13): 800-801. 10.1049/el:20080522 |
30 | WANG Z, BOVIK A C, SHEIKH H R, et al. Image quality assessment: from error visibility to structural similarity [J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612. 10.1109/tip.2003.819861 |
31 | ZHANG R, ISOLA P, EFROS A A, et al. The unreasonable effectiveness of deep features as a perceptual metric [C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 586-595. 10.1109/cvpr.2018.00068 |
32 | MITTAL A, SOUNDARARAJAN R, BOVIK A C. Making a “completely blind” image quality analyzer [J]. IEEE Signal Processing Letters, 2012, 20(3): 209-212. 10.1109/lsp.2012.2227726 |
33 | LIU L, LIU B, HUANG H, et al. No-reference image quality assessment based on spatial and spectral entropies[J]. Signal Processing: Image Communication, 2014, 29(8): 856-863. 10.1016/j.image.2014.06.006 |
34 | ZHANG H, PATEL V M. Densely connected pyramid dehazing network [C]// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 3194-3203. 10.1109/cvpr.2018.00337 |
35 | 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. 10.1109/tip.2016.2598681 |
36 | CHEN D, HE M, FAN Q, et al. Gated context aggregation network for image dehazing and deraining [C]// Proceedings of the 2019 IEEE Winter Conference on Applications of Computer Vision. Piscataway: IEEE, 2019: 1375-1383. 10.1109/wacv.2019.00151 |
37 | JU M, DING C, REN W, et al. IDE: image dehazing and exposure using an enhanced atmospheric scattering model [J]. IEEE Transactions on Image Processing, 2021, 30: 2180-2192. 10.1109/tip.2021.3050643 |
[1] | 赵志强, 马培红, 黑新宏. 基于双重注意力机制的人群计数方法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2886-2892. |
[2] | 秦璟, 秦志光, 李发礼, 彭悦恒. 基于概率稀疏自注意力神经网络的重性抑郁疾患诊断[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2970-2974. |
[3] | 李力铤, 华蓓, 贺若舟, 徐况. 基于解耦注意力机制的多变量时序预测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2732-2738. |
[4] | 贾洁茹, 杨建超, 张硕蕊, 闫涛, 陈斌. 基于自蒸馏视觉Transformer的无监督行人重识别[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2893-2902. |
[5] | 薛凯鹏, 徐涛, 廖春节. 融合自监督和多层交叉注意力的多模态情感分析网络[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2387-2392. |
[6] | 汪雨晴, 朱广丽, 段文杰, 李书羽, 周若彤. 基于交互注意力机制的心理咨询文本情感分类模型[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2393-2399. |
[7] | 高鹏淇, 黄鹤鸣, 樊永红. 融合坐标与多头注意力机制的交互语音情感识别[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2400-2406. |
[8] | 李钟华, 白云起, 王雪津, 黄雷雷, 林初俊, 廖诗宇. 基于图像增强的低照度人脸检测[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2588-2594. |
[9] | 莫尚斌, 王文君, 董凌, 高盛祥, 余正涛. 基于多路信息聚合协同解码的单通道语音增强[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2611-2617. |
[10] | 刘丽, 侯海金, 王安红, 张涛. 基于多尺度注意力的生成式信息隐藏算法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2102-2109. |
[11] | 徐松, 张文博, 王一帆. 基于时空信息的轻量视频显著性目标检测网络[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2192-2199. |
[12] | 李大海, 王忠华, 王振东. 结合空间域和频域信息的双分支低光照图像增强网络[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2175-2182. |
[13] | 魏文亮, 王阳萍, 岳彪, 王安政, 张哲. 基于光照权重分配和注意力的红外与可见光图像融合深度学习模型[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2183-2191. |
[14] | 熊武, 曹从军, 宋雪芳, 邵云龙, 王旭升. 基于多尺度混合域注意力机制的笔迹鉴别方法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2225-2232. |
[15] | 李欢欢, 黄添强, 丁雪梅, 罗海峰, 黄丽清. 基于多尺度时空图卷积网络的交通出行需求预测[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2065-2072. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||