生成對抗式網路(Generative Adversarial Network，GAN)[1] 是目前人工智慧最熱門的研究之一。GAN是一種強大的生成模型，其想法源自於博弈論的二人零和博弈，由一個生成器和一個判別器所組成，並透過對抗式學習的方式來達到訓練的效果。在以CNN為主軸的影像辨識工作上，在實務上遇到的第一個困難就是，如何蒐集大量的影像以供深度學習網路的訓練及測試使用。在虹膜辨識的演算法開發上，也有類似的問題。這篇論文中，我們提出了一種新的條件式生成對抗式網路，我們結合了WGAN-GP與一個獨立的分類器，使其能達到我們想要結果。運用此研究方法，可以根據我們自行設定的條件，產生某些特殊的影像，以此解決在進行深度學習實驗時訓練資料影像不足的問題，讓實驗達到更好的結果。

Generative Adversarial Network (GAN) is one of the most popular researches in the field of artificial intelligence. GAN is a powerful generation model. The idea is de-rived from the two-person zero-sum game of game theory. It consists of a generator and a discriminator. By simultaneously training these two models via adversarial net, both will become more powerful for the task they are designed to achieve. In the work of image identification based on CNN, the first difficulty in practice is how to collect enough images for the training and testing of the deep learning network. There are similar problems in the development of iris recognition algorithm. We construct a WGAN-GP combined with independent classifier, to achieve the de-sired results. Using this method, we can generate special images according to our condi-tions to solve the problem of insufficient image of training data in the course of deep learning experiments, and therefore, enhance the final recognition accuracy for the de-sired tasks.

[1] Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David WardeFarley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Advances in Neural Information Processing Systems 27, pages 2672–2680. Curran Associates, Inc., 2014.
[2]Bastien, F., Lamblin, P., Pascanu, R., Bergstra, J., Goodfellow, I. J., Bergeron, A., Bouchard, N., and Bengio, Y. (2012). Theano: new features and speed improvements. Deep Learning and Unsupervised Feature Learning NIPS 2012 Workshop.
[3] T. Salimans, I. Goodfellow, W. Zaremba, V. Cheung, A. Radford, and X. Chen. Improved techniques for training gans. In Advances in Neural Information Processing Systems, pages 2226–2234, 2016.
[4] Martin Arjovsky and L´eon Bottou. Towards principled methods for training generative adversarial networks. In International Conference on Learning Representations, 2017.
[5] M. Arjovsky, S. Chintala, and L. Bottou. Wasserstein gan. arXiv preprint arXiv:1701.07875, 2017.
[6] Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, and Aaron Courville. Improved training of Wasserstein GANs. arXiv preprint arXiv:1704.00028, 2017.
[7]Mirza, Mehdi and Osindero, Simon. Conditional generative adversarial nets. CoRR, abs/1411.1784, 2014.
[8] A. Odena, C. Olah, and J. Shlens. Conditional image synthesis with auxiliary classifier gans. arXiv preprint arXiv:1610.09585, 2016.
[9] D. Berthelot, T. Schumm, and L. Metz. Began: Boundary equilibrium generative adversarial networks. arXiv preprint arXiv:1703.10717, 2017
[10] X. Mao, Q. Li, H. Xie, R. Y. Lau, and Z. Wang. Least squares generative adversarial networks. arXiv preprint arXiv:1611.04076, 2016.
[11] Chen, X., Duan, Y., Houthooft, R., Schulman, J., Sutskever, I., and Abbeel, P. InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets. ArXiv e-prints, June 2016.
[12] David Berthelot, Tom Schumm, and Luke Metz. BEGAN: Boundary equilibrium generative adversarial networks. arXiv preprint arXiv:1703.10717, 2017.
[13] Mario Lucic, Karol Kurach, Marcin Michalski, Sylvain Gelly, Olivier Bousquet. Are GANs Created Equal? A Large-Scale Study. arXiv preprint arXiv:1711.10337,2017.
[14] Alec Radford, Luke Metz, and Soumith Chintala. Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv preprint arXiv:1511.06434, 2015.
[15] Yandong Wen, Kaipeng Zhangm, Zhifeng Li, Yu Qiao. A Discriminative Feature Learning Approach for Deep Face Recognition. Computer Vision – ECCV 2016,page 499-515, 2016.
[16] Sun, Y., Wang, X., Tang, X.: Deeply learned face representations are sparse, selective, and robust. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2892–2900 (2015)
[17] Schroff, F., Kalenichenko, D., Philbin, J.: Facenet: a unified embedding for face recognition and clustering. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 815–823 (2015)
[18] Parkhi, O.M., Vedaldi, A., Zisserman, A.: Deep face recognition. In: Proceedings of the British Machine Vision, vol. 1, no. 3, p. 6 (2015)
[19] Taigman, Y., Yang, M., Ranzato, M., Wolf, L.: Deepface: closing the gap to human-level performance in face verification. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1701–1708 (2014)
[20] S. Bell, C. L. Zitnick, K. Bala, and R. Girshick. Insideoutside net: Detecting objects in context with skip pooling and recurrent neural networks. In CVPR, 2016.
[21] T.-Y. Lin, P. Dollar, R. Girshick, K. He, B. Hariharan, and ´ S. Belongie. Feature pyramid networks for object detection. In CVPR, 2017
[22] R. Girshick. Fast R-CNN. In ICCV, 2015
[23] S. Ren, K. He, R. Girshick, and J. Sun. Faster R-CNN: Towards real-time object detection with region proposal networks. In NIPS, 2015.
[24] K. He, G. Gkioxari, P. Dollar, and R. B. Girshick. Mask ´ R-CNN. CoRR, abs/1703.06870, 2017
[25]http://biometrics.idealtest.org/dbDetailForUser.do?id=4.CAISA-IrisV4 database.