.png){kind=icon}
.png){kind=icon}
Return
Conditional image generation with One-Vs-All classifier
.png){kind=icon}
OA
AI
DOI：10.1016/j.neucom.2020.12.091.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper explores conditional image generation with a One-Vs-All classifier based on the Generative Adversarial Networks (GANs). Instead of the real/fake discriminator used in vanilla GANs, we propose to extend the discriminator to a One-Vs-All classifier (GAN-OVA) that can distinguish each input data to its category label. Specifically, we feed certain additional information as conditions to the generator and take the discriminator as a One-Vs-All classifier to identify each conditional category. Our model can be applied to different divergence or distances used to define the objective function, such as Jensen-Shannon divergence and Earth-Mover (or called Wasserstein-1) distance. We evaluate GANOVAs on MNIST and CelebA-HQ datasets, and the experimental results show that GAN-OVAs improve generation quality and the stability of training. (c) 2021 Elsevier B.V. All rights reserved.
Keywords:
Conditional image generation
Generative Adversarial Networks
One-Vs-All classifier
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.