簡易檢索 / 詳目顯示

回結果列表
研究生： 江心安
Jiang, Xin-An
論文名稱： 一個基於循環對抗網路的陰影移除演算法
A Cycle Generative Adversarial Networks Based Shadow Removal Algorithm
指導教授： 戴顯權
Tai, Shen-Chang
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： 2018
畢業學年度： 107
語文別： 英文
論文頁數： 71
中文關鍵詞： 陰影移除 、自然影像 、生成對抗網路 、卷積神經網路
外文關鍵詞： shadow removal, natural images, generative adversarial network, convolutional neural networks
相關次數： 點閱：51  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

    • 自然影像和影片通常含有陰影，但是陰影可能會影響演算法的效能。當處理含有陰影的輸入影像時，為了提高演算法的效能，進行陰影移除是必要的。
    本論文提出了一種基於循環生成對抗網絡的陰影去除演算法。此方法有兩個循環，一個循環用來增加陰影影像的多樣性，以來協助陰影去除產生器，而另一個循環則注重於陰影去除。兩個鑑別器判別兩個產生器的輸出是否是真實影像。所提出的方法，僅需要陰影和非陰影影像進行訓練，並不需要陰影遮罩。實驗結果顯示，本論文所提出的方法，在比較方法當中，得到了最佳的均方根誤差。有些比較方法，甚至需要用陰影遮罩用來訓練，或者在測試階段需要使用者輔助。

    Natural images and videos generally contain shadows, but shadows might affect the performance of algorithms. A shadow removal process is needed to improve the performance of algorithms when inputting images with shadows.
    In this thesis, a shadow removal algorithm based on cycle generative adversarial network is proposed. It has two cycles. One cycle increases the diversity of shadow images to help the shadow removal generator, and the other cycle focuses on shadow removal. The two discriminators distinguish whether the outputs from two generators are real images. The proposed method only requires shadow and non-shadow images for training, so it does not require shadow masks. Experimental results show that the proposed method get the best root mean square error in the compared methods. Even, some compared methods require shadow masks for training, or need a user-aided testing stage.

    Contents Contents iv List of Tables vi List of Figures vii Chapter 1 Introduction 1 Chapter 2 Background and Related Works 4 2.1 Neural Networks 4 2.2 Convolutional Neural Networks 12 2.3 VGG-16 Network 17 2.4 Generative Adversarial Networks 19 2.5 Cycle-Consistent Adversarial Network 22 Chapter 3 The Proposed Algorithm 24 3.1 Proposed Network Architecture 27 3.2 Loss Function 31 3.2.1 Cycle consistent loss 31 3.2.2 Ground truth loss 31 3.2.3 Adversarial loss 32 3.2.4 Shadow mask loss 32 3.2.5 Perceptual loss 35 3.2.6 Total variance loss 36 3.2.7 Total loss function 38 3.3 Generator Architecture 40 3.4 Discriminator Architecture 43 Chapter 4 Experimental Results 45 4.1 Experimental Dataset 45 4.2 Parameter and Experimental Setting 47 4.3 Experimental Results of Simulated Images 48 4.4 Ablation Experimental Results 59 Chapter 5 Conclusion and Future Work 66 5.1 Conclusion 66 5.2 Future Work 66 References 67

    [1] T. F. Y. Vicente, L. Hou, C.-P. Yu, M. Hoai, and D. Samaras. Large-scale training of shadow detectors with noisily annotated shadow examples. In Proceedings of the European Conference on Computer Vision, 2016.
    [2] J. Long, E. Shelhamer, and T. Darrell. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
    [3] V. Nguyen, T. F. Yago Vicente, M. Zhao, M. Hoai, and D. Samaras. Shadow detection with conditional generative adversarial networks. In IEEE International Conference on Computer Vision (ICCV), pages 4510–4518, 2017.
    [4] 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.
    [5] Hieu Le, Tomas F. Yago Vicente, Vu Nguyen, Minh Hoai, Dimitris Samaras. A+D-Net: Shadow Detection with Adversarial Shadow Attenuation. In Proceedings of the European Conference on Computer Vision, 2018.
    [6] I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio. Generative adversarial nets. In Neural Information Processing Systems (NIPS), 2014.
    [7] G. D. Finlayson, M. S. Drew, and C. Lu. Entropy minimization for shadow removal. International Journal of Computer Vision (IJCV), 85(1):35–57, 2009.
    [8] Q. Yang, K.-H. Tan, and N. Ahuja. Shadow removal using bilateral filtering. IEEE Transactions on Image Processing (TIP), 21(10):4361–4368, 2012.
    [9] H. Gong and D. Cosker. Interactive shadow removal and ground truth for variable scene categories. In British Machine Vision Conference (BMVC). University of Bath, 2014.
    [10] L. Qu, J. Tian, S. He, Y. Tang, and R. W. Lau. Deshadownet: A multi-context embedding deep network for shadow removal. In IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
    [11] S. H. Khan, M. Bennamoun, F. Sohel, and R. Togneri. Automatic shadow detection and removal from a single image. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 38(3):431–446, 2016.
    [12] J. Wang, X. Li, and J. Yang. Stacked conditional generative adversarial networks for jointly learning shadow detection and shadow removal. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018.
    [13] J. Zhu, T. Park, P. Isola, and A. A. Efros. Unpaired image-to-image translation using cycle-consistent adversarial networks. In International Conference on Computer Vision (ICCV), to appear, 2017.
    [14] Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun; The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770-778
    [15] K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recognition. In International Conference on Learning Representations (ICLR), 2015.
    [16] C. Ledig, L. Theis, F. Huszar, J. Caballero, A. Cunningham, ´ A. Acosta, A. Aitken, A. Tejani, J. Totz, Z. Wang, et al. Photo-realistic single image super-resolution using a generative adversarial network. arXiv preprint arXiv:1609.04802, 2016.
    [17] P. Isola, J.-Y. Zhu, T. Zhou, and A. A. Efros. Image-to-image translation with conditional adversarial networks. In IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
    [18] D. Pathak, P. Krahenbuhl, J. Donahue, T. Darrell, and A. A. Efros. Context encoders: Feature learning by inpainting. In IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
    [19] H. Zhang, V. A. Sindagi, and V. M. Patel. Image de-raining using a conditional generative adversarial network. arXiv preprint arXiv:1701.05957, 2017
    [20] A. Shrivastava, T. Pfister, O. Tuzel, J. Susskind, W. Wang, and R. Webb. Learning from simulated and unsupervised images through adversarial training.The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
    [21] 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.
    [22] Mahendran, A., Vedaldi, A. Understanding deep image representations by inverting them. In Proceedings of the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2015.
    [23] R. Guo, Q. Dai, and D. Hoiem. Paired regions for shadow detection and removal. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 35(12):2956–2967, 2013.
    [24] M. Gryka, M. Terry, and G. J. Brostow. Learning to remove soft shadows. ACM Transactions on Graphics (TOG), 34(5):153, 2015.
    [25] Y. Vicente, F. Tomas, L. Hou, C.-P. Yu, M. Hoai, and D. Samaras. Large-scale training of shadow detectors with noisily-annotated shadow examples. In European Conference on Computer Vision (ECCV), pages 816–832. Springer, 2016.
    [26] R. Guo, Q. Dai, and D. Hoiem. Paired regions for shadow detection and removal. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 35(12):2956–2967, 2013.
    [27] M. Mirza and S. Osindero. Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784, 2014.
    [28] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," in Advances in neural information processing systems, 2012, pp. 1097-1105.
    [29] K. Fukushima. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biological cybernetics, 1980.
    [30] 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.
    [31] A. L. Maas, A. Y. Hannun, and A. Y. Ng, "Rectifier nonlinearities improve neural network acoustic models," in Proc. ICML, 2013, vol. 30, no. 1.
    [32] A. Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, 2012, pp. 1097-1105.
    [33] Z. Yi, H. Zhang, T. Gong, Tan, and M. Gong. Dualgan: Unsupervised dual learning for image-to-image translation. In In IEEE International Conference on Computer Vision (ICCV), 2017.
    [34] J.-Y. Zhu, R. Zhang, D. Pathak, T. Darrell, A. A. Efros, O. Wang, and E. Shechtman. Toward multimodal image-to-image translation. In Advances in Neural Information Processing Systems (NIPS), 2017
    [35] M. D. Zeiler and R. Fergus. Visualizing and understanding convolutional networks. In D. J. Fleet, T. Pajdla, B. Schiele, and T. Tuytelaars, editors, ECCV, volume 8689 of Lecture Notes in Computer Science, pages 818–833. Springer, 2014.
    [36] S. H. Khan, M. Bennamoun, F. Sohel, and R. Togneri. Automatic feature learning for robust shadow detection. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1939–1946, 2014.
    [37] Y. Vicente, F. Tomas, M. Hoai, and D. Samaras. Leave-oneout kernel optimization for shadow detection and removal. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), PP(99):1–1, 2017.
    [38] A. Odena, C. Olah, and J. Shlens. Conditional Image Synthesis With Auxiliary Classifier GANs. ArXiv e-prints, Oct. 2016.
    [39] H. Chang, J. Lu, F. Yu, and A. Finkelstein. PairedCycleGAN: Asymmetric style transfer for applying and removing makeup. In Proceedings of the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), Jun. 2018
    [40] 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：2023-12-12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