A Style-Based Generator Architecture for Generative Adversarial Networks
Tero Karras,Samuli Laine,Timo Aila +2 more
- pp 4396-4405
Reads0
Chats0
TLDR
This paper proposed an alternative generator architecture for GANs, borrowing from style transfer literature, which leads to an automatically learned, unsupervised separation of high-level attributes (e.g., pose and identity when trained on human faces) and stochastic variation in the generated images.Abstract:
We propose an alternative generator architecture for generative adversarial networks, borrowing from style transfer literature. The new architecture leads to an automatically learned, unsupervised separation of high-level attributes (e.g., pose and identity when trained on human faces) and stochastic variation in the generated images (e.g., freckles, hair), and it enables intuitive, scale-specific control of the synthesis. The new generator improves the state-of-the-art in terms of traditional distribution quality metrics, leads to demonstrably better interpolation properties, and also better disentangles the latent factors of variation. To quantify interpolation quality and disentanglement, we propose two new, automated methods that are applicable to any generator architecture. Finally, we introduce a new, highly varied and high-quality dataset of human faces.read more
Citations
More filters
Book ChapterDOI
Controllable Image Synthesis via SegVAE
TL;DR: In this paper, SegVAE is proposed to synthesize semantic maps in an iterative manner using conditional variational autoencoder and image-to-image translation model.
Posted Content
Spectral Distribution Aware Image Generation
Steffen Jung,Margret Keuper +1 more
TL;DR: This paper proposes to generate images according to the frequency distribution of the real data by employing a spectral discriminator, which is lightweight, modular and works stably with different commonly used GAN losses.
Posted Content
Learning Group Structure and Disentangled Representations of Dynamical Environments
TL;DR: A physics-inspired method is proposed that learns a representation of an environment structured around the transformations that generate its evolution that allows for accurate long-horizon predictions and demonstrates a correlation between the quality of predictions and disentanglement in the latent space.
Journal ArticleDOI
Generative adversarial networks in medical image segmentation: A review
Siyi Xun,Dengwang Li,Hui Zhu,Min Chen,Jianbo Wang,Jie Li,Chen Meirong,Bing Wu,Hua Zhang,Xiangfei Chai,Zekun Jiang,Yan Zhang,Pu Huang +12 more
TL;DR: In this article, the authors reviewed more than 120 GAN-based architectures for medical image segmentation that were published before September 2021 and categorized and summarized these papers according to the segmentation regions, imaging modality, and classification methods.
Proceedings Article
Denoising Diffusion Implicit Models
TL;DR: In this article, the authors propose denoising diffusion implicit models (DDIMs), a more efficient class of iterative implicit probabilistic models with the same training procedure as DDPMs.