J
Jun-Yan Zhu
Researcher at Adobe Systems
Publications - 101
Citations - 59863
Jun-Yan Zhu is an academic researcher from Adobe Systems. The author has contributed to research in topics: Computer science & Generative model. The author has an hindex of 49, co-authored 96 publications receiving 42462 citations. Previous affiliations of Jun-Yan Zhu include University of California & University of California, Berkeley.
Papers
More filters
Posted Content
Ensembling with Deep Generative Views
TL;DR: In this article, the authors use StyleGAN2 as the source of GAN-based augmentations and investigate this setup on classification tasks involving facial attributes, cat faces, and cars, and find that while test-time ensembling with GANbased augmented images can offer some small improvements, the remaining bottlenecks are the efficiency and accuracy of the GAN reconstructions, coupled with classifier sensitivities to artifacts in GANgenerated images.
Posted Content
Diverse Image Generation via Self-Conditioned GANs
TL;DR: In this article, a class-conditional generative adversarial network (GAN) is proposed to generate realistic and diverse images without using manually annotated class labels, instead, their model is conditional on labels automatically derived from clustering in the discriminator's feature space.
Proceedings Article
Editing Conditional Radiance Fields
TL;DR: In this article, the authors propose a method for propagating coarse 2D user scribbles to the 3D space to modify the color or shape of a local region, which can be used for editing the appearance and shape of real photographs.
Journal ArticleDOI
Guest Editorial: Generative Adversarial Networks for Computer Vision
Supplementary Materials for: Unsupervised Object Class Discovery via Saliency-Guided Multiple Class Learning
TL;DR: An algorithm for simultaneously localizing objects and discovering object classes via bottom-up (saliency-guided) multiple class learning (bMCL) is developed, and significant improvements over the existing methods for multi-class object discovery are observed.