STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Multimedia Technology & Recent Advancements Original Research

A Study of DCGAN-Based Generative Models for Anime Character Face Generation

  1. Aaditya Mittal Department of Computer Science and Engineering, JECRC University, Jaipur
  2. Harsh Bathija Department of Computer Science and Engineering, JECRC University, Jaipur
  3. Yamini Sharma Department of Computer Science and Engineering, JECRC University, Jaipur
  4. Shreya Agarwal Department of Computer Science and Engineering, JECRC University, Jaipur

Abstract

Artificial intelligence, or AI, has in recent years moved from simple rule-based systems to models that are now fully capable of creative content generation and are referred to as generative AI. One such approach for content generation, introduced in the year 2014, is called generative adversarial networks (GANs), which consists of training a generator to create fake content that tries to mimic real content as closely as possible and a discriminator that tries to differentiate whether the content is real or fake, both contending against each other. Although GANs did show great potential for content generation, they lacked any stable way to generate images since their shallow architecture led to high volatility in training, with poor convergence, and poor image quality. To overcome these challenges, deep convolutional generative adversarial networks (DCGANs) were introduced in the year 2015, which were built upon traditional GANs by including convolutional and batch normalization layers in their architecture that allowed for stable and spatially coherent training for image generation. This project will research deeply into the performance of DCGAN and its advanced versions, with further optimizations in its architecture to overcome the limitations of GANs in generating images. In this study, a DCGAN model was trained to generate anime character images, exploring the benefits of convolutional architectures that allow stable training of adversarial models to improve the generated output qualitatively.

Keywords

References (20)

  1. Goodfellow IJ, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, et al. Generative adversarial nets. In: Lee DD, Sugiyama M, Luxburg UV, Guyon I, Garnett R, editors. Advances in Neural Information Processing Systems. Vol. 27. Red Hook (NY): Curran Associates Inc.; 2014.
  2. Radford A, Metz L, Chintala S. Unsupervised representation learning with deep convolutional generative adversarial networks [preprint]. 2015. arXiv:1511.06434. doi:10.48550/arXiv.1511.06434.
  3. Arjovsky M, Chintala S, Bottou L. Wasserstein generative adversarial networks. In: Precup D, Teh YW, editors. Proceedings of the 34th International Conference on Machine Learning; 2017 Aug 6–11; Sydney, Australia. Proceedings of Machine Learning Research. Vol. 70. PMLR; 2017. p. 214–223.
  4. Salimans T, Goodfellow I, Zaremba W, Cheung V, Radford A, Chen X. Improved techniques for training GANs. In: Lee DD, Sugiyama M, Luxburg UV, Guyon I, Garnett R, editors. Advances in Neural Information Processing Systems. Vol. 29. Red Hook (NY): Curran Associates Inc.; 2016.
  5. Gulrajani I, Ahmed F, Arjovsky M, Dumoulin V, Courville AC. Improved training of Wasserstein GANs. In: Guyon I, Luxburg UV, Bengio S, Wallach H, Fergus R, Vishwanathan S, et al., editors. Advances in Neural Information Processing Systems. Vol. 30. Red Hook (NY): Curran Associates Inc.; 2017.
  6. Jin Y, Zhang J, Li M, Tian Y, Zhu H, Fang Z. Towards the automatic anime characters creation with generative adversarial networks [preprint]. 2017. arXiv:1708.05509. doi:10.48550/arXiv.1708.05509.
  7. Li B, Zhu Y, Wang Y, Lin CW, Ghanem B, Shen L. AniGAN: Style-Guided Generative Adversarial Networks for Unsupervised Anime Face Generation. IEEE Transactions on Multimedia. 2022;24:4077-4091. doi:10.1109/tmm.2021.3113786
  8. Zhu JY, Park T, Isola P, Efros AA. Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks. 2017 IEEE International Conference on Computer Vision (ICCV). 2017:2242-2251. doi:10.1109/iccv.2017.244
  9. Liu MY, Tuzel O. Coupled generative adversarial networks. In: Lee DD, Sugiyama M, Luxburg UV, Guyon I, Garnett R, editors. Advances in Neural Information Processing Systems. Vol. 29. Red Hook (NY): Curran Associates Inc.; 2016.
  10. Karras T, Laine S, Aila T. A Style-Based Generator Architecture for Generative Adversarial Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2021;43(12):4217-4228. doi:10.1109/tpami.2020.2970919
  11. Karras T, Laine S, Aittala M, Hellsten J, Lehtinen J, Aila T. Analyzing and Improving the Image Quality of StyleGAN. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2020:8107-8116. doi:10.1109/cvpr42600.2020.00813
  12. Ioffe S, Szegedy C. Batch normalization: Accelerating deep network training by reducing internal covariate shift [preprint]. 2015. arXiv:1502.03167. doi:10.48550/arXiv.1502.03167.
  13. Kingma DP, Ba J. Adam: A method for stochastic optimization [preprint]. 2014. arXiv:1412.6980. doi:10.48550/arXiv.1412.6980.
  14. Zhang H, Goodfellow I, Metaxas D, Odena A. Self-attention generative adversarial networks. In: Chaudhuri K, Salakhutdinov R, editors. Proceedings of the 36th International Conference on Machine Learning. PMLR; 2019. p. 7354–7363.
  15. Li X, Li B, Fang M, Huang R, Huang X. BaMSGAN: Self-Attention Generative Adversarial Network with Blur and Memory for Anime Face Generation. Mathematics. 2023;11(20):4401. doi:10.3390/math11204401
  16. Liu M, Li Q, Qin Z, Zhang G, Wan P, Zheng W. BlendGAN: Implicitly GAN blending for arbitrary stylized face generation. In: Advances in Neural Information Processing Systems. Vol. 34. Red Hook (NY): Curran Associates Inc.; 2021. p. 29710–29722.
  17. Chakraborty T, Reddy K S U, Naik SM, Panja M, Manvitha B. Ten years of generative adversarial nets (GANs): a survey of the state-of-the-art. Machine Learning: Science and Technology. 2024;5(1):011001. doi:10.1088/2632-2153/ad1f77
  18. Dubey SR, Singh SK. Transformer-Based Generative Adversarial Networks in Computer Vision: A Comprehensive Survey. IEEE Transactions on Artificial Intelligence. 2024;5(10):4851-4867. doi:10.1109/tai.2024.3404910
  19. Joshi D, Sonawane S, Bhat T, Jumbad V, Wasade P, Zod S. Anime face generation using DC-GANs. AIP Conference Proceedings. 2023;2981:020028. doi:10.1063/5.0182666
  20. Cao YJ, Jia LL, Chen YX, Lin N, Yang C, Zhang B, et al. Recent Advances of Generative Adversarial Networks in Computer Vision. IEEE Access. 2019;7:14985-15006. doi:10.1109/access.2018.2886814
Support