Research map / 3D vision / Face and shape
Bootstrapping Articulated 3D Reconstruction from 2D Image Collections
Jakub Zadrozny, Oisin Mac Aodha, Hakan Bilen
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 |
|---|---|---|---|
| use heavy-tailed Student's t camera sampling for the canonical initialization (ablation) | canonical initialization · Horse-Mixed | mean CD (cm): 10.04 -> 7.24 | Improved |
| add iterative refinement from image collection without fitting regularization (ablation) | BAT3R · Horse-Mixed | mean CD (cm): 7.24 -> 5.59 | Improved |
| add geometric regularization in mesh fitting (ablation) | BAT3R · Horse-Mixed | mean CD (cm): 5.59 -> 4.85 | Improved |
| add viewpoint augmentation when rendering fitted meshes (ablation) | BAT3R · Horse-Mixed | mean CD (cm): 4.85 -> 3.86 | Improved |
| render fitted meshes into synthetic images instead of pairing raw images with estimated 3D (ablation) | BAT3R · Horse-Mixed | mean CD (cm): 6.96 -> 3.86 | Improved |
| fit realistic articulations from images instead of random pose sampling (ablation) | BAT3R · Horse-Mixed | mean CD (cm): 5.22 -> 3.86 | Improved |
| train on real in-the-wild horse images instead of DualPM synthetic images (ablation) | BAT3R · Horse-Robust | RMS CD (cm): 5.10 -> 6.01 | Hurt |
| refine from canonical initialization using synthetic images | BAT3R · Horse-Robust | RMS CD (cm): 7.84 -> 5.10 | Improved |
| train on real images vs canonical initialization | BAT3R · Horse-Robust | RMS CD (cm): 7.84 -> 6.01 | Improved |
| compare with fully supervised DualPM | BAT3R (synt) vs DualPM · Animodel-Points Horse | RMS CD (cm): 4.30 vs 5.65 | Hurt |
| number of refinement iterations (ablation) | BAT3R · Horse-Mixed | mean CD / outlier ratio: first iteration large jump, then plateau | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SAM-3D — SAM 3D: 3Dfy Anything in Images (CVPR 2026) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- R^2-Art: Category-Level Articulation Pose Estimation from Single RGB Image via Cascade Render Strategy (AAAI 2025 · similarity 0.89)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.88)
- SINGAPO: Single Image Controlled Generation of Articulated Parts in Objects (ICLR 2025 · similarity 0.88)
- Real2Code: Reconstruct Articulated Objects via Code Generation (ICLR 2025 · similarity 0.87)
- REACTO: Reconstructing Articulated Objects from a Single Video (CVPR 2024 · similarity 0.87)
- DIPO: Dual-State Images Controlled Articulated Object Generation Powered by Diverse Data (NeurIPS 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.
