`

Timezone: »

 
Spotlight
Instance-Conditioned GAN
Arantxa Casanova · Marlene Careil · Jakob Verbeek · Michal Drozdzal · Adriana Romero Soriano

@ None

Generative Adversarial Networks (GANs) can generate near photo realistic images in narrow domains such as human faces. Yet, modeling complex distributions of datasets such as ImageNet and COCO-Stuff remains challenging in unconditional settings. In this paper, we take inspiration from kernel density estimation techniques and introduce a non-parametric approach to modeling distributions of complex datasets. We partition the data manifold into a mixture of overlapping neighborhoods described by a datapoint and its nearest neighbors, and introduce a model, called instance-conditioned GAN (IC-GAN), which learns the distribution around each datapoint. Experimental results on ImageNet and COCO-Stuff show that IC-GAN significantly improves over unconditional models and unsupervised data partitioning baselines. Moreover, we show that IC-GAN can effortlessly transfer to datasets not seen during training by simply changing the conditioning instances, and still generate realistic images. Finally, we extend IC-GAN to the class-conditional case and show semantically controllable generation and competitive quantitative results on ImageNet; while improving over BigGAN on ImageNet-LT. Code and trained models to reproduce the reported results are available at https://github.com/facebookresearch/ic_gan.

Author Information

Arantxa Casanova (Polytechnique Montreal / Mila)
Marlene Careil (Facebook)
Jakob Verbeek (INRIA)
Michal Drozdzal (FAIR)
Adriana Romero Soriano (Facebook AI Research)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors