Analyzing and Improving the Image Quality of StyleGAN

  title={Analyzing and Improving the Image Quality of StyleGAN},
  author={Tero Karras and S. Laine and Miika Aittala and Janne Hellsten and J. Lehtinen and Timo Aila},
  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  • Tero Karras, S. Laine, +3 authors Timo Aila
  • Published 2020
  • Computer Science, Engineering, Mathematics
  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • The style-based GAN architecture (StyleGAN) yields state-of-the-art results in data-driven unconditional generative image modeling. We expose and analyze several of its characteristic artifacts, and propose changes in both model architecture and training methods to address them. In particular, we redesign the generator normalization, revisit progressive growing, and regularize the generator to encourage good conditioning in the mapping from latent codes to images. In addition to improving image… CONTINUE READING
    287 Citations
    One Model to Reconstruct Them All: A Novel Way to Use the Stochastic Noise in StyleGAN
    DeshuffleGAN: A Self-Supervised GAN to Improve Structure Learning
    • 3
    • PDF
    StyleGAN2 Distillation for Feed-forward Image Manipulation
    • 7
    • Highly Influenced
    • PDF
    Encoding in Style: a StyleGAN Encoder for Image-to-Image Translation
    • 8
    • PDF
    Style Intervention: How to Achieve Spatial Disentanglement with Style-based Generators?
    Unsupervised Image-to-Image Translation via Pre-trained StyleGAN2 Network
    Editing in Style: Uncovering the Local Semantics of GANs
    • 12
    • Highly Influenced
    • PDF
    In-Domain GAN Inversion for Real Image Editing
    • 20
    • PDF
    Improved Consistency Regularization for GANs
    • 8
    • PDF
    A U-Net Based Discriminator for Generative Adversarial Networks
    • 6
    • PDF


    Style Generator Inversion for Image Enhancement and Animation
    • 9
    • PDF
    Large Scale GAN Training for High Fidelity Natural Image Synthesis
    • 1,310
    • PDF
    A Style-Based Generator Architecture for Generative Adversarial Networks
    • 1,322
    • PDF
    MSG-GAN: Multi-Scale Gradient GAN for Stable Image Synthesis
    • 18
    • Highly Influential
    StackGAN: Text to Photo-Realistic Image Synthesis with Stacked Generative Adversarial Networks
    • 1,277
    • PDF
    Self-Attention Generative Adversarial Networks
    • 1,006
    • PDF
    On the Effects of Batch and Weight Normalization in Generative Adversarial Networks
    • 45
    StackGAN++: Realistic Image Synthesis with Stacked Generative Adversarial Networks
    • Han Zhang, T. Xu, +4 authors D. Metaxas
    • Computer Science, Medicine
    • IEEE Transactions on Pattern Analysis and Machine Intelligence
    • 2019
    • 340
    • PDF