在影像處理中，圖像去模糊是非常重要的，例如消除手震動時產生的模糊。在拍照時，我們經常取得模糊的圖像。為了解決此問題，我們提出了一種生成對抗網絡（deblur generative adversarial network with criss-cross attention, DGAN-CCA）來生成清晰的圖像。我們在網絡中加入縱橫交錯注意模塊，透過此機制讓網路能得到圖片中各點間資訊的依賴關係，藉此協助網路產生更加清晰的影像。我們還在訓練時使用光譜正規化 (spectral normalization)，透過此機制讓模型訓練更順利。縱橫交錯注意模塊幫助DGAN-CCA在PSNR和SSIM上相較本來的方法有更好的表現。

Image deblurring is very important in image processing such as removing image blur caused by hand shaking. When we take pictures, we often get blurry images. To solve the problem, we propose a deblur generative adversarial network with criss-cross attention (DGAN-CCA) to generate a sharp image without blurring artifacts. We add a criss-cross attention module to help the network obtain feature dependency between different image areas for deblurring. We also utilize spectral normalization to train the network. The criss-cross attention module helps DGAN-CCA perform better against the baseline in terms of peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM).

PyTorch, http://pytorch.org.
O. Kupyn, V. Budzan, M. Mykhailych, D. Mishkin, and J. Matas. “Deblurgan: Blind motion deblurring using conditional adversarial networks.” CVPR, pp. 8183–8192, 2018
O. Kupyn, T. Martyniuk, J.Wu, and Z. Wang. DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), pp. 16–20, 2019.
I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio. “Generative Adversarial Networks.”, 2014.
J. Cheng, L. Dong, and M. Lapata. “Long short-term memory-networks for machine reading.”, EMNLP, 2016.
AP. Parikh, O. Täckström, D. Das, J. Uszkoreit. “A decomposable attention model for natural language inference.” EMNLP, 2016.
X. Wang, R. Girshick, A. Gupta, and K. He.” Non-local neural networks.” CVPR, 2018.
H. Zhang, I. Goodfellow, D. Metaxas, and A. Odena. “Self-attention generative adversarial networks.” PMLR, 2019.
P. Isola, JY. Zhu, T. Zhou, and AA. Efros. “Image-to-image translation with conditional adversarial networks.” CVPR, 2017.
R. Szeliski. “Computer Vision: Algorithms and Applications.” USA, 1st edition, 2010.
J. Sun, W. Cao, Z. Xu, and J. Ponce. “Learning a Convolutional Neural Network for Non-uniform Motion Blur Removal.”, 2015.
D. Gong, J. Yang, L. Liu, Y. Zhang, I. Reid, C. Shen, AVD. Hengel, and Q. Shi. “From Motion Blur to Motion Flow: a Deep Learning Solution for Removing Heterogeneous Motion Blur.” CVPR, 2017.
X. Tao, H. Gao, X. Shen, J. Wang, and J. Jia. “Scale-recurrent network for deep image deblurring.” CVPR, 2018
S. Nah, T. Hyun, K. Kyoung, and M. Lee. “Deep Multiscale Convolutional Neural Network for Dynamic Scene Deblurring.” CVPR, 2017.
S. Ramakrishnan, S. Pachori, A. Gangopadhyay, and S. Raman. “Deep generative filter for motion deblurring.” ICCVW, pp. 2993–3000, 2017.
D. Kingma and J. Ba. “Adam: A method for stochastic optimization.” ICLR, 2014.
M. Arjovsky, S. Chintala, and L. Bottou. “Wasserstein gan.” arXiv preprint arXiv:1701.07875, 2017.
I. Gulrajani, F. Ahmed, M. Arjovsky, V. Dumoulin, and AC. Courville. “Improved training of wasserstein gans.” Advances in neural information processing systems, pp. 5767–5777, 2017.
TY. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie. “Feature pyramid networks for object detection.” CVPR, pp. 2117–2125, 2017.
D. Bahdanau, K. Cho, and Y. Bengio. “Neural machine translation by jointly learning to align and translate.” arXiv:1409.0473, 2014.
K. Xu, J. Ba, R. Kiros, K. Cho, AC. Courville, R. Salakhutdinov, RS. Zemel, and Y. Bengio. “Show, attend and tell: Neural image caption generation with visual attention.” ICML, 2015.
Z. Yang, X. He, J. Gao, L. Deng, and AJ. Smola. “Stacked attention networks for image question answering.” CVPR, 2016.
K. Gregor, I. Danihelka,A. Graves, DJ. Rezende, and D. Wierstra. “DRAW: A recurrent neural network for image generation.” ICML, 2015.
X. Chen, N. Mishra, M. Rohaninejad, and P. Abbeel. “Pixelsnail: An improved autoregressive generative model.” ICML, 2018.
A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, AN. Gomez, L. Kaiser, and I. Polosukhin. “Attention is all you need.” NIPS, 2017.
Z. Huang, X. Wang, L. Huang, C. Huang, Y. Wei, and W. Liu. “Ccnet: Criss-cross attention for semantic segmentation.” ICCV, 2019.
T. Miyato, T. Kataoka, M. Koyama, and Y. Yoshida. “Spectral Normalization for Generative Adversarial Networks.” ICLR, 2018.
S. Nah, T. Hyun, K. Kyoung, and M. Lee. “Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurring.” CPVR, 2017.
HK. Galoogahi, A. Fagg, C. Huang, D. Ramanan, and S. Lucey. “Need for speed: A benchmark for higher frame rate object tracking.” ICCV, pp. 1125–1134, 2017.
S. Su, M. Delbracio, J. Wang, G. Sapiro, W. Heidrich, and O. Wang. “Deep video deblurring for hand-held cameras.” CVPR, 2017.
X. Mao, Q. Li, H. Xie, R. Y.K. Lau, Z. Wang, S. P. Smolley. “Least Squares Generative Adversarial Networks.” CVPR, 2017.
WS. Lai, JB. Huang, Z. Hu, N. Ahuja, MH. Yang. “A Comparative Study for Single Image Blind Deblurring.” CVPR, 2016.