Learning Generative Models of Textured 3D Meshes from Real-World Images

TL;DR

This paper introduces a GAN-based method for generating textured 3D meshes from images without needing pose annotations, achieving comparable results to prior methods and extending to more diverse categories like ImageNet.

Contribution

The authors develop a pose-annotation-free GAN framework for textured 3D mesh generation, broadening applicability to datasets lacking keypoint annotations.

Findings

Achieves performance comparable to keypoint-based methods.

Sets new baselines on diverse ImageNet categories.

Operates without class-specific hyperparameter tuning.

Abstract

Recent advances in differentiable rendering have sparked an interest in learning generative models of textured 3D meshes from image collections. These models natively disentangle pose and appearance, enable downstream applications in computer graphics, and improve the ability of generative models to understand the concept of image formation. Although there has been prior work on learning such models from collections of 2D images, these approaches require a delicate pose estimation step that exploits annotated keypoints, thereby restricting their applicability to a few specific datasets. In this work, we propose a GAN framework for generating textured triangle meshes without relying on such annotations. We show that the performance of our approach is on par with prior work that relies on ground-truth keypoints, and more importantly, we demonstrate the generality of our method by setting new baselines on a larger set of categories from ImageNet - for which keypoints are not available - without any class-specific hyperparameter tuning. We release our code at https://github.com/dariopavllo/textured-3d-gan