Learning Generative Models of Textured 3D Meshes from Real-World Images
This work addresses the limitation of requiring delicate pose estimation with annotated keypoints, enabling broader applicability in computer graphics and generative modeling for real-world images.
The paper tackles the problem of learning generative models of textured 3D meshes from image collections without relying on annotated keypoints, achieving performance on par with prior methods that use ground-truth keypoints and setting new baselines on a larger set of ImageNet categories.
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