Compositional GAN: Learning Conditional Image Composition
By Samaneh Azadi, Deepak Pathak, Sayna Ebrahimi, Trevor Darrell. 2018
26 citations recorded by the source corpus.
Generative Adversarial Networks (GANs) can produce images of surprising complexity and realism, but are generally modeled to sample from a single latent source ignoring the explicit spatial interaction between multiple entities that could be present in a scene. Capturing such complex interactions between different objects in the world, including their relative scaling, spatial layout, occlusion, or viewpoint transformation is a challenging problem. In this work, we propose to model object composition in a GAN framework as a self-consistent composition-decomposition network. Our model is conditioned on the object images from their marginal distributions to generate a realistic image from their joint distribution by explicitly learning the possible interactions. We evaluate our model through qualitative experiments and user evaluations in both the scenarios when either paired or unpaired examples for the individual object images and the joint scenes are given during training. Our results reveal that the learned model captures potential interactions between the two object domains given as input to output new instances of composed scene at test time in a reasonable fashion.
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Trevor Darrell
- Sergey Levine
- Sjoerd van Steenkiste
- Sayna Ebrahimi
- Yoshua Bengio
- Marcus Rohrbach
- Abhinav Kumar Gupta
- Jitendra Malik
Citations in the corpus
- Adversarial Continual Learning (2020, methodology): “ey are often limited by the number of tasks. 4 S. Ebrahimi et al. 2.2 Adversarial learning Adversarial learning has been used for different problems such as generative models [12], object composition [2], representation learning [21], domain adaptation [35], active learning [33], etc. The use of an ”
- A Case for Object Compositionality in Deep Generative Models of Images (2018, background): “Azadi et al. (2018) considers a framework to generate images composed of two objects, conditioned on images of each single object.”