CDL-GAN: Contrastive Distance Learning Generative Adversarial Network for Image Generation
Source abstract
Abstract
While Generative Adversarial Networks (GANs) have shown promising performance in image generation, they suffer from numerous issues such as mode collapse and training instability. To stabilize GAN training and improve image synthesis quality with diversity, we propose a simple yet effective approach as Contrastive Distance Learning GAN (CDL-GAN) in this paper. Specifically, we add Consistent Contrastive Distance (CoCD) and Characteristic Contrastive Distance (ChCD) into a principled framework to improve GAN performance. The CoCD explicitly maximizes the ratio of the distance between generated images and the increment between noise vectors to strengthen image feature learning for the generator. The ChCD measures the sampling distance of the encoded images in Euler space to boost feature representations for the discriminator. We model the framework by employing Siamese Network as a module into GANs without any modification on the backbone. Both qualitative and quantitative experiments conducted on three public datasets demonstrate the effectiveness of our method.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Generative Adversarial Networks and Image Synthesis · Advanced Image Processing Techniques · Image Processing Techniques and Applications
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Computer Engineering and Science, Shanghai University, Shanghai 200444, China
Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, China
Funding
National Natural Science Foundation of China · National Key Research and Development Program of China
References (48)
Show source-provided references
1. Goodfellow, I.J., Pouget-Abadie, J., Mirza, M., Bing, X., and Bengio, Y. (2014, January 8–13). Generative adversarial nets. Proceedings of the Advances in Neural Information Processing Systems, Montreal, QC, Canada.
2. Mirza, M., and Osindero, S. (2014). Conditional generative adversarial nets. arXiv.
3. Huang, X., Li, Y., Poursaeed, O., Hopcroft, J., and Belongie, S. (2017, January 21–26). Stacked generative adversarial networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA.10.1109/CVPR.2017.202 ↗
4. Odena, A., Olah, C., and Shlens, J. (2017, January 6–11). Conditional image synthesis with auxiliary classifier gans. Proceedings of the International Conference on Machine Learning, Sydney, Australia.
5. Wang, Y., Zhang, L., and Van De Weijer, J. (2016). Ensembles of generative adversarial networks. arXiv.
6. Hoang, Q., Nguyen, T.D., Le, T., and Phung, D. (2017). Multi-generator generative adversarial nets. arXiv.
7. Nguyen, T., Le, T., Vu, H., and Phung, D. (2017, January 4–9). Dual discriminator generative adversarial nets. Proceedings of the Advances in Neural Information Processing Systems, Long Beach, CA, USA.
8. Larsen, A.B.L., Sønderby, S.K., Larochelle, H., and Winther, O. (2016, January 19–24). Autoencoding beyond pixels using a learned similarity metric. Proceedings of the International Conference on Machine Learning, New York, NY, USA.
9. Makhzani, A., Shlens, J., Jaitly, N., Goodfellow, I., and Frey, B. (2015). Adversarial autoencoders. arXiv.
10. Dumoulin, V., Belghazi, I., Poole, B., Mastropietro, O., Lamb, A., Arjovsky, M., and Courville, A. (2016). Adversarially learned inference. arXiv.
11. Wang, X., and Wang, X. (2018). Unsupervised Domain Adaptation with Coupled Generative Adversarial Autoencoders. Appl. Sci., 8.10.3390/app8122529 ↗
12. Kwak, J.g., and Ko, H. (2020). Unsupervised Generation and Synthesis of Facial Images via an Auto-Encoder-Based Deep Generative Adversarial Network. Appl. Sci., 10.10.3390/app10061995 ↗
13. Arjovsky, M., Chintala, S., and Bottou, L. (2017). Wasserstein gan. arXiv.
14. Miyato, T., Kataoka, T., Koyama, M., and Yoshida, Y. (2018). Spectral normalization for generative adversarial networks. arXiv.
15. Salimans, T., Zhang, H., Radford, A., and Metaxas, D. (2018). Improving GANs using optimal transport. arXiv.
16. Li, C.L., Chang, W.C., Cheng, Y., Yang, Y., and Póczos, B. (2017, January 4–9). Mmd gan: Towards deeper understanding of moment matching network. Proceedings of the Advances in Neural Information Processing Systems, Long Beach, CA, USA.
17. DeVries, T., and Taylor, G.W. (2017). Improved regularization of convolutional neural networks with cutout. arXiv.
18. Wei, X., Gong, B., Liu, Z., Lu, W., and Wang, L. (2018). Improving the improved training of wasserstein gans: A consistency term and its dual effect. arXiv.
19. Odena, A., Zhang, H., Lee, H., and Zhang, Z. (2019). Consistency Regularization for Generative Adversarial Networks. arXiv.10.23915/distill.00018 ↗
20. Mao, Q., Lee, H.Y., Tseng, H.Y., Ma, S., and Yang, M.H. (2019, January 16–20). Mode seeking generative adversarial networks for diverse image synthesis. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA.10.1109/CVPR.2019.00152 ↗
21. Zhao, Z., Singh, S., Lee, H., Zhang, Z., Odena, A., and Zhang, H. (2020). Improved consistency regularization for gans. arXiv.10.1609/aaai.v35i12.17317 ↗
22. Ansari, A.F., Scarlett, J., and Soh, H. (2020, January 13–19). A Characteristic Function Approach to Deep Implicit Generative Modeling. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.
23. Mescheder, L., Nowozin, S., and Geiger, A. (2017, January 4–9). The numerics of gans. Proceedings of the Advances in Neural Information Processing Systems, Long Beach, CA, USA.
24. Daskalakis, C., Ilyas, A., Syrgkanis, V., and Zeng, H. (2017). Training gans with optimism. arXiv.
25. Prasad, H., LA, P., and Bhatnagar, S. (2015, January 4–8). Two-timescale algorithms for learning Nash equilibria in general-sum stochastic games. Proceedings of the 2015 International Conference on Autonomous Agents and Multiagent Systems, Istanbul, Turkey.
26. Yadav, A., Shah, S., Xu, Z., Jacobs, D., and Goldstein, T. (2017). Stabilizing adversarial nets with prediction methods. arXiv.
27. Miyato, T., and Koyama, M. (2018). cGANs with projection discriminator. arXiv.
28. Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., and Hochreiter, S. (2017, January 4–9). Gans trained by a two time-scale update rule converge to a local nash equilibrium. Proceedings of the Advances in Neural Information Processing Systems, Long Beach, CA, USA.
29. Lim, J.H., and Ye, J.C. (2017). Geometric gan. arXiv.
30. Villani, C. (2008). Optimal Transport: Old and New, Springer Science & Business Media.
31. Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., and Courville, A.C. (2017, January 4–9). Improved training of wasserstein gans. Proceedings of the Advances in Neural Information Processing Systems, Long Beach, CA, USA.
32. Virtual adversarial training: A regularization method for supervised and semi-supervised learning10.1109/TPAMI.2018.2858821 ↗
33. Chwialkowski, K.P., Ramdas, A., Sejdinovic, D., and Gretton, A. (2015, January 7–12). Fast two-sample testing with analytic representations of probability measures. Proceedings of the Advances in Neural Information Processing Systems, Montreal, QC, Canada.
34. An omnibus test for the two-sample problem using the empirical characteristic function10.1080/00949658608810963 ↗
35. A test of goodness of fit for symmetric random variables110.1111/j.1467-842X.1972.tb00355.x ↗
36. Bromley, J., Guyon, I., Lecun, Y., Sckinger, E., and Shah, R. (1993, January 22–23). Signature Verification Using a Siamese Time Delay Neural Network. Proceedings of the Advances in Neural Information Processing Systems 6, 7th NIPS Conference, Denver, CO, USA.
37. Nair, V., and Hinton, G.E. (2010, January 21–24). Rectified linear units improve restricted boltzmann machines. Proceedings of the 27th International Conference on Machine Learning, Haifa, Israel.
38. Bertinetto, L., Valmadre, J., Henriques, J.F., Vedaldi, A., and Torr, P.H.S. (October, January 8–10,). Fully-Convolutional Siamese Networks for Object Tracking. Proceedings of the Computer Vision—ECCV 2016 Workshops, Amsterdam, The Netherlands.10.1007/978-3-319-48881-3_56 ↗
39. Xu, Z., Luo, H., Hui, B., Chang, Z., and Ju, M. (2019). Siamese Tracking with Adaptive Template-Updating Strategy. Appl. Sci., 9.10.3390/app9183725 ↗
40. Amodio, M., and Krishnaswamy, S. (2019, January 16–20). Travelgan: Image-to-image translation by transformation vector learning. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA.10.1109/CVPR.2019.00919 ↗
41. Zagoruyko, S., and Komodakis, N. (2015, January 7–12). Learning to compare image patches via convolutional neural networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, USA.10.1109/CVPR.2015.7299064 ↗
42. Da, K. (2014). A method for stochastic optimization. arXiv.
43. Radford, A., Metz, L., and Chintala, S. (2015). Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv.
44. Gradient-based learning applied to document recognition10.1109/5.726791 ↗
45. Krizhevsky, A., and Hinton, G. (2009). Learning multiple layers of features from tiny images. Handbook of Systemic Autoimmune Diseases, Elsevier.
46. Liu, Z., Luo, P., Wang, X., and Tang, X. (2015, January 7–13). Deep Learning Face Attributes in the Wild. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Santiago, Chile.10.1109/ICCV.2015.425 ↗
47. Kurach, K., Lucic, M., Zhai, X., Michalski, M., and Gelly, S. (2019). A Large-Scale Study on Regularization and Normalization in GANs. arXiv.
48. Metz, L., Poole, B., Pfau, D., and Sohl-Dickstein, J. (2016). Unrolled Generative Adversarial Networks. arXiv.