Research map / 3D vision / Face and shape
DOVE: Learning Deformable 3D Objects by Watching Videos
Shangzhe Wu, Tomáš Jakab, Christian Rupprecht, Andrea Vedaldi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 front-back hypothesis (two-view ambiguity resolution) (ablation) | DOVE · 3D Toy Bird Scans | Chamfer Distance (cm): 2.52 -> 1.51 | Improved |
| add symmetry constraint (ablation) | DOVE · 3D Toy Bird Scans | Chamfer Distance (cm): 2.19 -> 1.51 | Improved |
| add video training (ablation) | DOVE · 3D Toy Bird Scans | Chamfer Distance (cm): 2.20 -> 1.51 | Improved |
| add learned category prior shape (ablation) | DOVE · 3D Toy Bird Scans | Chamfer Distance (cm): 3.92 -> 1.51 | Improved |
| train on mask-only supervision instead of extra geometric supervision | CMR · 3D Toy Bird Scans | Chamfer Distance (cm): 1.35 -> 5.94 | Hurt |
| remove template shape and keypoints for baselines | VMR · 3D Toy Bird Scans | Chamfer Distance (cm): 1.28 -> 1.90 | Hurt |
| use DOVE with mask, flow and video only vs UMR mask only | UMR · 3D Toy Bird Scans | Chamfer Distance (cm): 2.26 -> 1.51 | Improved |
| add SCOPS part segmentation to UMR | UMR · Bird video mIoU dt=0 | mIoU: 0.853 -> 0.830 | Hurt |
| transfer articulation from frame t+dt | DOVE · Bird video mIoU dt=20 | mIoU: 0.805 -> 0.845 | Improved |
| use finetuned CMR with extra supervision vs DOVE | CMR · 3D Toy Bird Scans | Chamfer Distance (cm): 1.35 -> 1.51 | Mixed |
Closest papers by meaning
- BANMo: Building Animatable 3D Neural Models From Many Casual Videos (CVPR 2022 · similarity 0.88)
- SPAMs: Structured Implicit Parametric Models (CVPR 2022 · similarity 0.88)
- Unsupervised Learning of Dense Shape Correspondence (CVPR 2019 · similarity 0.88)
- Learning Category-Specific Mesh Reconstruction from Image Collections (ECCV 2018 · similarity 0.88)
- Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses (NeurIPS 2024 · similarity 0.87)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.87)
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.
