Research map / 3D vision / Face and shape
Unsupervised Shape and Pose Disentanglement for 3D Meshes
Keyang Zhou, Bharat Lal Bhatnagar, Gerard Pons-Moll
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 self-consistency loss to cross-consistency (ablation) | Ours · AMASS pose transfer | mean error (mm): 23.83 -> 19.43 | Improved |
| add ARAP deformation in the training loop (ablation) | Ours · AMASS pose transfer | mean error (mm): 20.27 -> 19.43 | Improved |
| replace pose supervision with self-consistency | Ours supervised · AMASS pose transfer | mean error (mm): 15.44 -> 19.43 | Mixed |
| use unsupervised mesh disentanglement instead of GDVAE | GDVAE · AMASS pose transfer | mean error (mm): 54.44 -> 19.43 | Improved |
| keep full latent dimension instead of PCA reduction | Ours · AMASS shape retrieval (shape code) | E_beta: 0.34 -> 0.14 | Improved |
| reduce latent dimension with PCA for pose retrieval | Ours · AMASS pose retrieval (pose code) | E_beta: 0.92 -> 2.14 | Mixed |
| use Ours instead of Jiang et al. on COMA | Jiang et al. · COMA expression extrapolation | average error (mm): 1.64 -> 1.28 | Improved |
| add edge length constraints and other local shape-preserving losses (ablation) | Ours · AMASS | reconstruction | No clear effect |
Closest papers by meaning
- Unsupervised Part-Based Disentangling of Object Shape and Appearance (CVPR 2019 · similarity 0.90)
- NPMs: Neural Parametric Models for 3D Deformable Shapes (ICCV 2021 · similarity 0.89)
- Unsupervised 3D Pose Transfer With Cross Consistency and Dual Reconstruction (TPAMI 2023 · similarity 0.89)
- Disentangling Geometric Deformation Spaces in Generative Latent Shape Models (IJCV 2023 · similarity 0.89)
- Unsupervised Learning of Dense Shape Correspondence (CVPR 2019 · similarity 0.89)
- Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses (NeurIPS 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.
