Diffusion self-guidance for controllable image generation

D Epstein, A Jabri, B Poole, A Efros… - Advances in Neural …, 2023 - proceedings.neurips.cc
Large-scale generative models are capable of producing high-quality images from detailed
prompts. However, many aspects of an image are difficult or impossible to convey through …

Survey on leveraging pre-trained generative adversarial networks for image editing and restoration

M Liu, Y Wei, X Wu, W Zuo, L Zhang - Science China Information Sciences, 2023 - Springer
Generative adversarial networks (GANs) have drawn enormous attention due to their simple
yet effective training mechanism and superior image generation quality. With the ability to …

Blobgan: Spatially disentangled scene representations

D Epstein, T Park, R Zhang, E Shechtman… - European Conference on …, 2022 - Springer
We propose an unsupervised, mid-level representation for a generative model of scenes.
The representation is mid-level in that it is neither per-pixel nor per-image; rather, scenes …

Linkgan: Linking gan latents to pixels for controllable image synthesis

J Zhu, C Yang, Y Shen, Z Shi, B Dai… - Proceedings of the …, 2023 - openaccess.thecvf.com
This work presents an easy-to-use regularizer for GAN training, which helps explicitly link
some axes of the latent space to a set of pixels in the synthesized image. Establishing such …

Householder projector for unsupervised latent semantics discovery

Y Song, J Zhang, N Sebe… - Proceedings of the IEEE …, 2023 - openaccess.thecvf.com
Abstract Generative Adversarial Networks (GANs), especially the recent style-based
generators (StyleGANs), have versatile semantics in the structured latent space. Latent …

High-fidelity gan inversion with padding space

Q Bai, Y Xu, J Zhu, W Xia, Y Yang, Y Shen - European Conference on …, 2022 - Springer
Abstract Inverting a Generative Adversarial Network (GAN) facilitates a wide range of image
editing tasks using pre-trained generators. Existing methods typically employ the latent …

Flow factorized representation learning

Y Song, A Keller, N Sebe… - Advances in Neural …, 2024 - proceedings.neurips.cc
A prominent goal of representation learning research is to achieve representations which
are factorized in a useful manner with respect to the ground truth factors of variation. The …

Bilinear models of parts and appearances in generative adversarial networks

J Oldfield, C Tzelepis, Y Panagakis… - … on Pattern Analysis …, 2024 - ieeexplore.ieee.org
Recent advances in the understanding of Generative Adversarial Networks (GANs) have led
to remarkable progress in visual editing and synthesis tasks, capitalizing on the rich …

Panda: Unsupervised learning of parts and appearances in the feature maps of gans

J Oldfield, C Tzelepis, Y Panagakis, A Nicolaou… - 2023 - qmro.qmul.ac.uk
Recent advances in the understanding of Generative Adversarial Networks (GANs) have led
to remarkable progress in visual editing and synthesis tasks, capitalizing on the rich …

Exploring sparse moe in gans for text-conditioned image synthesis

J Zhu, C Yang, K Zheng, Y Xu, Z Shi, Y Shen - arXiv preprint arXiv …, 2023 - arxiv.org
Due to the difficulty in scaling up, generative adversarial networks (GANs) seem to be falling
from grace on the task of text-conditioned image synthesis. Sparsely-activated mixture-of …