结合FC-DenseNet和WGAN的图像去雾算法

doi:10.3778/j.issn.1673-9418.1905044

摘要/Abstract

摘要：

针对现有图像去雾算法严重依赖中间量准确估计的问题，提出了一种基于Wasserstein生成对抗网络（WGAN）的端到端图像去雾模型。首先，使用全卷积密集块网络（FC-DenseNet）充分学习图像中雾的特征；其次，采用残差学习思想直接从退化图像中学习到清晰图像的特征，实现端到端的去雾；最后，使用均方误差和感知结构误差函数作为模型的损失函数，以确保生成图像结构和内容的相似度，并使用WGAN对生成结果细致优化，生成清晰逼真的无雾图像。实验结果表明，在合成雾天数据集上，该算法在结构相似度上比其他对比算法提高了4%；在自然雾天图像上，由该算法恢复的图像具有较高的清晰度和对比度，在主观评价上优于其他对比算法。

关键词: 图像去雾, Wasserstein生成对抗网络（WGAN）, 全卷积密集块网络（FC-DenseNet）, 残差学习

Abstract:

The existing image dehazing algorithms rely heavily on the accurate estimation of the intermediate variables. This paper proposes an end-to-end image dehazing model based on Wasserstein generative adversarial networks(WGAN). Firstly, the fully convolutional DenseNets (FC-DenseNet) is used to fully learn the features of the hazy in image. Secondly, the residual learning concept is used to directly learn the features of the clear image from the degraded image, and realize end-to-end image dehazing. Finally, the mean square error and perceptual structural error function are used as the loss function of the model to ensure the image structure and content information, and WGAN is used to finely optimize the generated results to produce clear and realistic clear images. Experimental results show that the proposed algorithm improves the structural similarity by 4% compared with other comparison algorithms on the synthetic hazy dataset, and on the natural hazy image, the image restored by the algorithm has higher definition and contrast, and is superior to other comparison algorithms on the subjective evaluation.

Key words: image dehazing, Wasserstein generative adversarial networks (WGAN), fully convolutional DenseNet(FC-DenseNet), residual learning

孙斌，雎青青，桑庆兵. 结合FC-DenseNet和WGAN的图像去雾算法[J]. 计算机科学与探索, 2020, 14(8): 1380-1388.

SUN Bin, JU Qingqing, SANG Qingbing. Image Dehazing Algorithm Based on FC-DenseNet and WGAN[J]. Journal of Frontiers of Computer Science and Technology, 2020, 14(8): 1380-1388.

参考文献

[1] Tan R T. Visibility in bad weather from a single image[C]// Proceedings of the 2008 IEEE Conference on Computer Vision and Pattern Recognition, Anchorage, Jun 23-28, 2008. Washington: IEEE Computer Society, 2008: 1-8.
[2] He K M, Sun J, Tang X O. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341-2353.
[3] Zhu Q S, Mai J M, Shao L. A fast single image haze removal algorithm using color attenuation prior[J]. IEEE Transactions on Image Processing, 2015, 24(11): 3522-3533.
[4] Cai B L, Xu X M, 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.
[5] Ren W Q, Liu S, Zhang H, et al. Single image dehazing via multi-scale convolutional neural networks[C]//LNCS 9906: Proceedings of the 14th European Conference on Computer Vision, Amsterdam, Oct 8-10, 2016. Berlin, Heidelberg: Sprin-ger, 2016: 154-169.
[6] Li B Y, Peng X L, Wang Z Y, et al. AOD-Net: all-in-one dehazing network[C]//Proceedings of the 2017 IEEE Inter-national Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 4770-4778.
[7] McCartney E J. Optics of the atmosphere: scattering by mole-cules and particles[M]. New York: John Wiley & Sons, Inc., 1976.
[8] Galdran A, Vazquezcorral J, Pardo D, et al. Fusion-based varia-tional image dehazing[J]. IEEE Signal Processing Letters, 2017, 24(2): 151-155.
[9] Zhang D, Wu P. Fast single image defogging algorithm[J]. Computer Engineering and Applications, 2019, 55(10): 213-217.张弟, 吴萍. 单幅图像快速去雾算法[J]. 计算机工程与应用, 2019, 55(10): 213-217.
[10] Arjovsky M, Chintala S, Bottou L. Wasserstein GAN[C]//Pro-ceedings of the 2017 International Conference on Machine Learning, Sydney, Aug 6-11, 2017. New York: ACM, 2017: 214-223.
[11] Jégou S, Drozdzal M, Vázquez D, et al. The one hundred layers tiramisu: fully convolutional densenets for semantic segmentation[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 1175-1183.
[12] Goodfellow I J, Pouget A J, Mirza M, et al. Generative adver-sarial networks[J]. Advances in Neural Information Processing Systems, 2014, 3: 2672-2680.
[13] Ledig C, Theis L, Huszar F, et al. Photo-realistic single image super-resolution using a generative adversarial network[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 105-114.
[14] Isola P, Zhu J Y, Zhou T H, et al. Image-to-image translation with conditional adversarial networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 5967-5976.
[15] Kupyn O, Budzan V, Mykhailych M, et al. DeblurGAN: blind motion deblurring using conditional adversarial networks[C]//Proceedings of the 2018 IEEE/CVF Conference on Com-puter Vision and Pattern Recognition, Salt Lake City, Jun 18-22, 2018. Washington: IEEE Computer Society, 2018: 8183-8192.
[16] Ioffe S, Szegedy C. Batch normalization: accelerating deep network training by reducing internal covariate shift[C]// Proceedings of the 32nd International Conference on Machine Learning, Lille, Jul 6-11, 2015: 448-456.
[17] Ronneberger O, Fischer P, Brox T. U-Net: convolutional net-works for biomedical image segmentation[C]//LNCS 9351: Proceedings of the 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, Mu-nich, Oct 5-9, 2015. Berlin, Heidelberg: Springer, 2015: 234-241.
[18] Huang G, Liu Z, Laurens V D M, et al. Densely connected convolutional networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 4700-4708.
[19] He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 26-Jul 1, 2016. Washington: IEEE Computer Society, 2016: 770-778.
[20] Mirza M, Osindero S. Conditional generative adversarial nets[J]. arXiv:1411.1784, 2014.
[21] Ding Y, Huang L, Wang C D. Link prediction based on generative adversarial networks[J]. Journal of Frontiers of Computer Science and Technology, 2019, 13(4): 554-562.丁玥, 黄玲, 王昌栋. 基于生成式对抗网络的链路预测方法[J]. 计算机科学与探索, 2019, 13(4): 554-562.
[22] Johnson J, Alahi A, Li F F. Perceptual losses for real-time style transfer and super-resolution[C]//LNCS 9906: Proceed-ings of the 14th European Conference on Computer Vision, Amsterdam, Oct 11-14, 2016. Berlin, Heidelberg: Springer, 2016: 694-711.
[23] Ju Q Q, Li C F, Sang Q B. Single image dehazing by using improved multi-scale convolutional neural network[J]. Com-puter Engineering and Applications, 2019, 55(10): 179-185.雎青青, 李朝锋, 桑庆兵. 改进多尺度卷积神经网络的单幅图像去雾方法[J]. 计算机工程与应用, 2019, 55(10): 179-185.
[24] Silberman N, Hoiem D, Kohli P, et al. Indoor segmentation and support inference from RGBD images[C]//LNCS 7576: Proceedings of the 12th European Conference on Computer Vision, Firenze, Oct 7-13, 2012. Berlin, Heidelberg: Springer, 2012: 746-760.
[25] Wang Y J, Li J H, Lu Y, et al. Image quality evaluation based on image weighted separating block peak signal to noise ratio[C]//Proceedings of the 2003 International Conference on Neural Networks and Signal Processing, Nanjing, Dec 14-17, 2003. Piscataway: IEEE, 2003: 994-997.
[26] Wang Z, Bovik A C, Sheikh H R, et al. Image quality assess-ment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612.