Research map / 3D vision / Face and shape
Unsupervised Learning of Probably Symmetric Deformable 3D Objects From Images in the Wild
Check 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 albedo flipping to enforce canonical symmetry (ablation) | Ours on BFM · BFM | SIDE (×10^-2): 2.916 -> 0.793 | Helped |
| Add depth flipping to enforce canonical symmetry (ablation) | Ours on BFM · BFM | SIDE (×10^-2): 1.139 -> 0.793 | Helped |
| Compute shading from predicted depth and light instead of predicting a shading map (ablation) | Ours on BFM · BFM | SIDE (×10^-2): 2.406 -> 0.793 | Helped |
| Add perceptual loss (ablation) | Ours on BFM · BFM | SIDE (×10^-2): 0.931 -> 0.793 | Helped |
| Replace the ImageNet-pretrained perceptual encoder with a self-supervised encoder (ablation) | Ours on BFM · BFM | SIDE (×10^-2): 0.793 -> 0.815 | No clear effect |
| Add learned confidence maps (ablation) | Ours on BFM · BFM | SIDE (×10^-2): 0.829 -> 0.793 | No clear effect |
| Add learned confidence maps to handle asymmetric image perturbations (ablation) | Ours on perturbed BFM · BFM with asymmetric perturbations | SIDE (×10^-2): 2.141 -> 0.878 | Helped |
| Use CelebA pre-training | Ours on 3DFAW · 3DFAW frontal faces | Depth Corr.: 48.98 -> 54.65 | Helped |
Closest papers by meaning
- Modelling and unsupervised learning of symmetric deformable object categories (NeurIPS 2018 · similarity 0.93)
- Canonical 3D Deformer Maps: Unifying parametric and non-parametric methods for dense weakly-supervised category reconstruction (NeurIPS 2020 · similarity 0.90)
- Symmetric Shape-Preserving Autoencoder for Unsupervised Real Scene Point Cloud Completion (CVPR 2023 · similarity 0.88)
- Unsupervised Part-Based Disentangling of Object Shape and Appearance (CVPR 2019 · similarity 0.88)
- Unsupervised Learning of Category-Level 3D Pose from Object-Centric Videos (CVPR 2024 · similarity 0.87)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.86)
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.