3inGAN: Learning a 3D Generative Model from Images of a Self-similar Scene

TL;DR

3inGAN is a novel 3D generative model trained from a single 2D image of a self-similar scene, capable of producing view-consistent 3D scene variations and remixes without flickering artifacts.

Contribution

It introduces a method to learn plausible, view-consistent 3D scene variations from only one exemplar image, combining 2D and 3D GAN losses for realism.

Findings

Successfully generates view-consistent 3D scene variations

Works with real and synthetic scenes of varying complexity

Outperforms recent related methods in qualitative and quantitative metrics

Abstract

We introduce 3inGAN, an unconditional 3D generative model trained from 2D images of a single self-similar 3D scene. Such a model can be used to produce 3D "remixes" of a given scene, by mapping spatial latent codes into a 3D volumetric representation, which can subsequently be rendered from arbitrary views using physically based volume rendering. By construction, the generated scenes remain view-consistent across arbitrary camera configurations, without any flickering or spatio-temporal artifacts. During training, we employ a combination of 2D, obtained through differentiable volume tracing, and 3D Generative Adversarial Network (GAN) losses, across multiple scales, enforcing realism on both its 3D structure and the 2D renderings. We show results on semi-stochastic scenes of varying scale and complexity, obtained from real and synthetic sources. We demonstrate, for the first time, the feasibility of learning plausible view-consistent 3D scene variations from a single exemplar scene and provide qualitative and quantitative comparisons against recent related methods.