Research map / 3D vision / Face and shape
DensePose 3D: Lifting Canonical Surface Maps of Articulated Objects to the Third Dimension
Roman Shapovalov, David Novotny, Benjamin Graham, Patrick Labatut, 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 canonicalisation loss (ablation) | DP3D · UP-3D | MPJPE: 183.6 -> 91.2 | Improved |
| add canonicalisation loss (ablation) | DP3D · Human3.6M (Protocol 1) | RE14: 135.4 -> 113.6 | Improved |
| add canonicalisation loss (ablation) | DP3D · 3DPW (transfer from H3.6M) | RE14: 120.3 -> 95.2 | Improved |
| add canonicalisation loss (ablation) | DP3D · Stanford Dogs | MPJPE: 241.4 -> 247.1 | Hurt |
| add ARAP loss (ablation) | DP3D · UP-3D | MPJPE: 242.6 -> 91.2 | Improved |
| add ARAP loss (ablation) | DP3D · Human3.6M (Protocol 1) | RE14: 154.8 -> 113.6 | Improved |
| add ARAP loss (ablation) | DP3D · Stanford Dogs | MPJPE: 371.8 -> 247.1 | Improved |
| add entropy loss on part assignment (ablation) | DP3D · UP-3D | MPJPE: 113.8 -> 91.2 | Improved |
| add entropy loss on part assignment (ablation) | DP3D · 3DPW (transfer from H3.6M) | RE14: 99.1 -> 95.2 | Improved |
| add entropy loss on part assignment (ablation) | DP3D · Stanford Dogs | MPJPE: 505.2 -> 247.1 | Improved |
| use latent part-based articulation instead of linear LBO blendshapes (ablation) | DP3D · UP-3D | MPJPE: 205.9 -> 91.2 | Improved |
| use latent part-based articulation instead of linear LBO blendshapes (ablation) | DP3D · Human3.6M (Protocol 1) | RE14: 125.0 -> 113.6 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- C3DPO — C3DPO: Canonical 3D Pose Networks for Non-Rigid Structure From Motion (ICCV 2019) · better on 4 of 4 shared results
- SPIN — SPIN: Structure-Preserving Inner Offset Network for Scene Text Recognition (AAAI 2021) · better on 0 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
- Learning Monocular 3D Reconstruction of Articulated Categories From Motion (CVPR 2021 · similarity 0.91)
- Canonical 3D Deformer Maps: Unifying parametric and non-parametric methods for dense weakly-supervised category reconstruction (NeurIPS 2020 · similarity 0.90)
- Learning Articulated Shape With Keypoint Pseudo-Labels From Web Images (CVPR 2023 · similarity 0.90)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.88)
- SAOR: Single-View Articulated Object Reconstruction (CVPR 2024 · similarity 0.88)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 2024 · 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.
