Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling

TL;DR

This paper introduces 3D-GAN, a novel generative adversarial network that creates high-quality 3D objects from a probabilistic space, enabling unsupervised shape learning and recognition.

Contribution

The paper presents a new 3D-GAN framework that captures object structure implicitly and learns shape features without supervision, advancing 3D object generation and recognition.

Findings

Generates high-quality 3D objects

Learns effective 3D shape features unsupervised

Achieves recognition performance comparable to supervised methods

Abstract

We study the problem of 3D object generation. We propose a novel framework, namely 3D Generative Adversarial Network (3D-GAN), which generates 3D objects from a probabilistic space by leveraging recent advances in volumetric convolutional networks and generative adversarial nets. The benefits of our model are three-fold: first, the use of an adversarial criterion, instead of traditional heuristic criteria, enables the generator to capture object structure implicitly and to synthesize high-quality 3D objects; second, the generator establishes a mapping from a low-dimensional probabilistic space to the space of 3D objects, so that we can sample objects without a reference image or CAD models, and explore the 3D object manifold; third, the adversarial discriminator provides a powerful 3D shape descriptor which, learned without supervision, has wide applications in 3D object recognition. Experiments demonstrate that our method generates high-quality 3D objects, and our unsupervisedly learned features achieve impressive performance on 3D object recognition, comparable with those of supervised learning methods.