Timezone: »

 
Poster
RELATE: Physically Plausible Multi-Object Scene Synthesis Using Structured Latent Spaces
Sebastien Ehrhardt · Oliver Groth · Aron Monszpart · Martin Engelcke · Ingmar Posner · Niloy Mitra · Andrea Vedaldi

Tue Dec 08 09:00 AM -- 11:00 AM (PST) @ Poster Session 1 #220
in Poster Session 2 »

We present RELATE, a model that learns to generate physically plausible scenes and videos of multiple interacting objects. Similar to other generative approaches, RELATE is trained end-to-end on raw, unlabeled data. RELATE combines an object-centric GAN formulation with a model that explicitly accounts for correlations between individual objects. This allows the model to generate realistic scenes and videos from a physically-interpretable parameterization. Furthermore, we show that modeling the object correlation is necessary to learn to disentangle object positions and identity. We find that RELATE is also amenable to physically realistic scene editing and that it significantly outperforms prior art in object-centric scene generation in both synthetic (CLEVR, ShapeStacks) and real-world data (cars). In addition, in contrast to state-of-the-art methods in object-centric generative modeling, RELATE also extends naturally to dynamic scenes and generates videos of high visual fidelity. Source code, datasets and more results are available at http://geometry.cs.ucl.ac.uk/projects/2020/relate/.

Author Information

Sebastien Ehrhardt (University of Oxford)
Oliver Groth (University of Oxford)
Aron Monszpart (Niantic)
Martin Engelcke (University of Oxford)
Ingmar Posner (Oxford University)
Niloy Mitra (UCL/Adobe)
Andrea Vedaldi (Facebook AI Research and University of Oxford)

More from the Same Authors