簡易檢索 / 詳目顯示

回結果列表
研究生： 黃子育
Huang, Tzu-Yu
論文名稱： 使用循環對抗網路的單張影像除霧
Single image dehazing using cycle consistent adversarial networks
指導教授： 戴顯權
Tai, Shen-Chuan
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 電腦與通信工程研究所
Institute of Computer & Communication Engineering
論文出版年： 2019
畢業學年度： 107
語文別： 英文
論文頁數： 58
中文關鍵詞： 影像除霧 、循環生成對抗網路 、自然影像 、深度學習
外文關鍵詞： image dehazing, CycleGAN, natural image, deep learning
相關次數： 點閱：89  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 電腦視覺運用的相當廣泛，像是監視系統、影像辨識等等都有可能利用到。然而在影像辨識領域，影像品質大大的影響了其判斷的結果與準確性。戶外所拍攝出來的影像很大程度會受到空氣中的懸浮微粒影響。光線不走直線進行散射，使得拍攝出來的影像夾帶霧氣。有霧的影像會大幅度影響影像品質，非常不利於影像辨識且造成判斷上準確率的下降。近年來有許多學者提出單張影像的除霧方法，致力於解決此窘境。本篇論文中，我們利用深度學習的方式進行除霧。基於循環生成對抗網路，同時加上暗通道與深度資訊等等條件，使我們在數值上與視覺上，比其他現有除霧方法的實驗結果要優。

    Computer vision is widely used in many fields, such as surveillance system, image recognition, etc. However, image quality highly affects the result and accuracy in image recognition. Images that taken at outdoor are highly affected by particles in air. Particles cause light scattering, so the images become hazy. Hazy images will highly decrease the image quality. It is not conducive to image recognition and it will decrease the accuracy. In recent years, many researchers proposed many single image dehazing methods aim to solve the problem. In this Thesis, a method based on cycle consistent adversarial networks is proposed. After adding the dark channel, depth map and other losses into the networks, our experiment has a better result both in visual and quantitative metrics.

    Contents iv List of Tables vi List of Figures vii Chapter 1 Introduction 1 Chapter 2 Background and Related Works 3 2.1 Theory of Light Propagation 3 2.2 Dark Channel Prior 5 2.2.1 Dark Channel Prior 5 2.2.2 Estimating the Transmission 7 2.2.3 Estimating the Atmospheric Light 8 2.2.4 Soft matting 10 2.2.5 Recovering the Scene Radiances 11 2.3 Deep Learning 12 2.3.1 Neural Network 13 2.3.2 Convolutional Neural Network 17 2.3.3 Cycle Consistent Adversarial Networks 19 2.4 Cycle Dehaze 21 Chapter 3 The Proposed Algorithm 23 3.1 Dehazing cycle and Haze-adding cycle 27 3.2 Generator Architecture 30 3.3 Discriminator Architecture 30 3.4 Loss Function 33 3.4.1 Adversarial loss 33 3.4.2 Cycle consistency loss 34 3.4.3 Ground truth loss 35 3.4.4 Perceptual loss 36 3.4.5 Dark channel loss 38 3.4.6 Depth loss 39 3.4.7 Total loss function 39 Chapter 4 Experimental Results 41 4.1 Experimental Dataset 41 4.2 Ablation Experimental Results 43 4.3 Parameters and Specifications 44 4.4 Experimental Results 45 Chapter 5 Conclusions and Future Works 55 5.1 Conclusions 55 5.2 Future Works 55 References 56

    [1] S. G. Narasimhan, S. K. Nayar, “Chromatic framework for vision in bad weather,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 1, pp.598-605, 2000.
    [2] S. G. Narasimhan and S. K. Nayar, “Contrast restoration of weather degraded images,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 25, no. 6, pp.713-724, Jun. 2003.
    [3] S. Shwartz, E. Namer, and Y.Y. Schechner, “Blind Haze Separation,” IEEE Conf. Computer Vision and Pattern Recognition, vol. 2, pp. 1984-1991, 2006.
    [4] Y. Y. Schechner, S. G. Narasimhan, and S. K. Nayar, “Instant Dehazing of Images Using Polarization,” IEEE Conf, Computer Vision and Pattern Recognition, vol.1, pp. 325-332, 2001.
    [5] R. Tan, “Visibility in Bad Weather from a Single Image,” Proc. IEEE Conf. Computer Vision and Pattern Recognition, June 2008.
    [6] K. He, J. Sun, and X. Tang, “Single image haze removal using dark channel prior,” in Proc. IEEE CVPR, pp. 1956-1963, June 2009.
    [7] I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D.Warde-Farley, S.Ozair, A.Courville, and Y.Bengio, “Generative adversarial nets,” in Advances in Neural Information Processing Systems, pp. 2672-2680, 2014.
    [8] J. Y. Zhu, T. Park, P. Isola, and A. A. Efros, “Unpaired image-to-image translation using cycle-consistent adversarial networks,” IEEE International Conference on Computer Vision, 2017
    [9] N. Silberman, D. Hoiem, P. Kohli, and R. Fergus, “Indoor segmentation and support inference from rgbd images,” in European Conference on Computer Vision, pp. 746-760. Springer, 2012.
    [10] C. Ancuti, C. O. Ancuti, and C. De Vleeschouwer, “D-HAZY: a dataset to evaluate quantitatively dehazing algorithms,” in IEEE international Conference on Image Processing, pp. 2226-2230, 2016.
    [11] H. Koschmieder, “Theorie der Horizontalen Sichtweite,” in Beitrage zur Physik der Freien Atmosphare. Munich, Germany: Keim & Nemnich, vol. 12, pp. 171-181, 1924.
    [12] E. J. McCartney, “Optics of the Atomosphere: Scattering by Molecules and Particles,” New York, John Wiley and Sons, 1976.
    [13] J. Kopf, B.Neubert, B.Chen, M.Cohen, D. Cohen-Or, O.Deussen, M. Uyttendaele, and D. Lischinski, “Deep photo: model-based photograph enhancement and viewing,” ACM Transactions on Graphics, vol.27, no. 5, pp. 1-10, Dec. 2008.
    [14] K. Fukushima, “Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position,” Biological cybernetics, 1980.
    [15] Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, and L. D. Jackel, “Backpropagation applied to handwritten zip code recognition,” Neural computation, 1989.
    [16] D. Engin, A. Genc, H. K. Ekenel, “Cycle-Dehaze: Enhanced CycleGAN for Single Image Dehazing,” The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2018.
    [17] K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” In International Conference on Learning Representations (ICLR), 2015.
    [18] O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” In International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pages 234–241. Springer, 2015.
    [19] Q. Zhu, J. Mai, and L. Shao, “A fast single image haze removal algorithm using color attenuation prior,” IEEE Transactions on Image Processing, 24(11):3522-3533, 2015.
    [20] J. Johnson, A. Alahi, and L. Fei-Fei, “Perceptual losses for real-time style transfer and super-resolution,” In Proceedings of the European Conference on Computer Vision (ECCV), pages 694–711. Springer, 2016.
    [21] Guilin Liu, Fitsum A Reda, Kevin J Shih, Ting-Chun Wang, Andrew Tao, and Bryan Catanzaro, “Image inpainting for irregular holes using partial convolutions,” arXiv preprint arXiv:1804.07723, 2018.

    無法下載圖示 校內：立即公開
    校外：不公開
    電子論文尚未授權公開，紙本請查館藏目錄
    QR CODE