Research map / 3D vision / Face and shape
Visual Object Networks: Image Generation with Disentangled 3D Representations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add a learned 3D shape prior and disentangled shape, viewpoint, and texture generation | VON · Car images | FID: 130.5 -> 81.6 | Helped |
| Add a learned 3D shape prior and disentangled shape, viewpoint, and texture generation | VON · Chair images | FID: 225.3 -> 58.0 | Helped |
| Add a learned 3D shape prior and disentangled shape, viewpoint, and texture generation | VON (DF) · Car images vs DCGAN, LSGAN, and WGAN-GP | Human preference for VON (DF): 72.2%; 78.7%; 63.0 % | Helped |
| Add a learned 3D shape prior and disentangled shape, viewpoint, and texture generation | VON (DF) · Chair images vs DCGAN, LSGAN, and WGAN-GP | Human preference for VON (DF): 90.3%; 92.4%; 89.1% | Helped |
| Use Wasserstein loss with gradient penalty instead of cross-entropy loss | 3D shape generator · ShapeNet cars and chairs | FID: 3.021 -> 0.021; 2.598 -> 0.082 | Helped |
| Use distance-function rather than voxel representation (ablation) | VON · Car images | FID: 81.6 -> 83.3 | Mixed |
| Use distance-function rather than voxel representation (ablation) | VON · Chair images | FID: 58.0 -> 51.8 | Helped |
| Use distance-function rather than voxel representation (ablation) | VON shape generator · ShapeNet cars and chairs (DF representation) | FID: 3.896 -> 0.002; 1.790 -> 0.006 | Mixed |
| Add a latent texture-code cycle-consistency loss | VON texture generator · 2.5D sketches to 2D images | Texture-code reconstruction | Helped |
Closest papers by meaning
- Leveraging 2D Data to Learn Textured 3D Mesh Generation (CVPR 2020 · similarity 0.91)
- Towards Unsupervised Learning of Generative Models for 3D Controllable Image Synthesis (CVPR 2020 · similarity 0.90)
- Synthesizing 3D Shapes via Modeling Multi-View Depth Maps and Silhouettes With Deep Generative Networks (CVPR 2017 · similarity 0.90)
- Disentangled3D: Learning a 3D Generative Model With Disentangled Geometry and Appearance From Monocular Images (CVPR 2022 · similarity 0.90)
- Convolutional Generation of Textured 3D Meshes (NeurIPS 2020 · similarity 0.90)
- Cue3D: Quantifying the Role of Image Cues in Single-Image 3D Generation (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.