基于条件生成对抗网络的图像去雾算法
Image Dehazing Algorithm Based on Conditional Generation against Network
摘要: 本文改进了一种基于条件生成对抗网络的图像去雾算法。在生成器网络模型中使用32层Tiramisu代替U-Net可以减少训练参数，提高参数利用率。判别器网络的最后一层使用Sigmoid函数完成归一化。使用损失函数加权和的方式训练网络模型，使用一种评价去雾能力的分数，保存训练过程中分数较高的模型及参数，最后使用分数最高的网络模型及参数对室外真实图像进行去雾测试。训练集使用了室内和室外的真实无雾图像以及合成有雾图像，测试结果显示本文提出的模型和传统去雾方法相比，主观视觉效果、图像细节、色彩方面都有提高改善。
Abstract: This paper improves an image dehazing algorithm based on conditional generation against networks. Using 32 Layer Tiramisu instead of U-Net in the generator network model can reduce training parameters and improve parameter utilization. The final layer of the discriminator network uses the Sigmoid function to complete the normalization. The network model is trained using the weighted sum of loss function, using a score to evaluate the ability of defogging, saving the model and parameters with higher scores during the training process, and finally using the network model with the highest score and parameters to perform the dehazing test on the outdoor real image. The training set uses real fog-free images and synthetic foggy images both indoors and outdoors. The test results show that the proposed model has improved subjective visual effects, image detail and color compared with the traditional dehazing method.
文章引用：姚远, 李向阳, 孟金慧. 基于条件生成对抗网络的图像去雾算法[J]. 图像与信号处理, 2020, 9(1): 1-7. https://doi.org/10.12677/JISP.2020.91001

参考文献

[1] Tan, R.T. (2008) Visibility in Bad Weather from a Single Image. IEEE Conference on Computer Vision and Pattern Recognition, Anchorage, 23-28 June 2008, 1-8.
[2] Kopf, J., Neubert, B., Chen, B., Cohen, M., Cohen-Or, D., Deussen, O., Uyttendaele, M. and Lischinski, D. (2008) Deep Photo: Model-Based Photograph Enhancement and Viewing. ACM Transactions on Graphics, 27, Article No. 116.
[Google Scholar] [CrossRef] 
[3] He, K.M., Sun, J. and Tang, X.O. (2010) Single Image Haze Removal Using Dark Channel Prior. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33, 2341-2353.
[Google Scholar] [CrossRef] 
[4] Meng, G.F., Wang, Y., Duan, J.Y., Xiang, S.M. and Pan, C.H. (2013) Efficient Image Dehazing with Boundary Constraint and Contextual Regularization. Proceedings of the IEEE International Conference on Computer Vision, Sydney, 1-8 December 2013, 617-624.
[Google Scholar] [CrossRef] 
[5] Berman, D., Avidan, S., et al. (2016) Non-Local Image Dehazing. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 26 June-1 July 2016, 1674-1682.
[Google Scholar] [CrossRef] 
[6] Cai, B., Xu, X.M., Jia, K., Qing, C.M. and Tao, D.C. (2016) Dehazenet: An End-to-End System for Single Image Haze Removal. IEEE Transactions on Image Processing, 25, 5187-5198.
[Google Scholar] [CrossRef] 
[7] Ren, W.Q., Liu, S., Zhang, H., Pan, J.S., Cao, X.C. and Yang, M.-H. (2016) Single Image Dehazing via Multi-Scale Convolutional Neural Networks. In: European Conference on Computer Vision, Springer, Cham, 154-169.
[Google Scholar] [CrossRef] 
[8] Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A. and Bengio, Y. (2014) Generative Adversarial Nets. Advances in Neural Information Processing Systems, Montreal, 8-13 December 2014, 2672-2680.
[9] Isola, P., Zhu, J.-Y., Zhou, T.H. and Efros, A.A. (2017) Image-to-Image Translation with Conditional Adversarial Networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, 21-26 July 2017, 1125-1134.
[Google Scholar] [CrossRef] 
[10] Zhang, H., Sindagi, V. and Patel, V.M. (2017) Joint Transmission Map Estimation and Dehazing Using Deep Networks.
[11] McCartney, E.J. (1976) Optics of the Atmosphere: Scattering by Molecules and Particles. John Wiley and Sons, Inc., New York, 421.
[12] Narasimhan, S.G. and Nayar, S.K. (2003) Contrast Restoration of Weather Degraded Images. IEEE Transactions on Pattern Analysis & Machine Intelligence, 25, 713-724.
[Google Scholar] [CrossRef] 
[13] Mirza, M. and Osindero, S. (2014) Conditional Generative Adversarial Nets.
[14] Jégou, S., Drozdzal, M., Vazquez, D., Romero, A. and Bengio, Y. (2017) The One Hundred Layers Tiramisu: Fully Convolutional Densenets for Semantic Segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Honolulu, 21-26 July 2017, 11-19.
[Google Scholar] [CrossRef] 
[15] Radford, A., Metz, L. and Chintala, S. (2015) Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks.
[16] 梁毓明, 张路遥, 卢明建, 杨国亮. 基于条件生成对抗网络的图像去雾算法[J]. 光子学报, 2019, 48(5): 0510002.
[17] Simonyan, K. and Zisserman, A. (2014) Very Deep Convolutional Networks for Large-Scale Image Recognition.
[18] Silberman, P.K.N., Hoiem, D. and Fergus, R. (2012) Indoor Segmentation and Support Inference from Rgbd Images. In: European Conference on Computer Vision, Springer, Berlin, Heidelberg, 746-760.
[Google Scholar] [CrossRef] 
[19] Ancuti, C.O., Ancuti, C., Timofte, R. and De Vleeschouwer, C. (2018) O-Haze: A Dehazing Benchmark with Real Hazy and Haze-Free Outdoor Images. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Salt Lake City, 18-22 June 2018, 754-762.
[Google Scholar] [CrossRef] 
[20] Kingma, D.P. and Ba, J. (2014) Adam: A Method for Stochastic Optimization.