Research map / 3D vision / Face and shape
MHR-Net: Multiple-Hypothesis Reconstruction of Non-rigid Shapes from 2D Views
Haitian Zeng, Xin Yu, Jiaxu Miao, Yi Yang
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 basis-deformation shape representation without intermediate loss (ablation) | baseline NRSfM · Human3.6M (perspective) | MPJPE: 83.5 -> 198.3 | Hurt |
| add intermediate loss on the basis shape (ablation) | MHR-Net with basis-deformation · Human3.6M (perspective) | MPJPE: 198.3 -> 75.5 | Improved |
| add stochastic deformation with best hypothesis selection (DBSD) (ablation) | MHR-Net BD+IL · Human3.6M (perspective) | MPJPE: 75.5 -> 73.7 | Improved |
| use best instead of worst hypothesis for optimization (ablation) | MHR-Net DBSD · Human3.6M (perspective) | MPJPE: 78.4 -> 73.7 | Improved |
| use best hypothesis instead of mixture density (ablation) | MHR-Net DBSD · Human3.6M (perspective) | MPJPE: 77.9 -> 73.7 | Improved |
| add Procrustean Residual Loss (ablation) | MHR-Net DBSD best hypothesis · Human3.6M (perspective) | MPJPE: 73.7 -> 72.6 | Improved |
| add Procrustean Residual Loss (ablation) | MHR-Net DBSD worst hypothesis · Human3.6M (perspective) | MPJPE: 78.4 -> 76.2 | Improved |
| add Procrustean Residual Loss (ablation) | MHR-Net DBSD mixture density · Human3.6M (perspective) | MPJPE: 77.9 -> 77.3 | Improved |
| add Procrustean Residual Loss (ablation) | MHR-Net · SURREAL | normalized error: 0.1388 -> 0.1192 | Improved |
| compare worst hypothesis to best hypothesis at inference (N_m=50) | MHR-Net · Human3.6M (perspective) | MPJPE: -0.5 and -0.1 MPJPE at epoch 10 and epoch 50 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 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
- PAUL — PAUL: Procrustean Autoencoder for Unsupervised Lifting (CVPR 2021) · 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
- Deep Non-Rigid Structure From Motion With Missing Data (TPAMI 2020 · similarity 0.88)
- Structure From Recurrent Motion: From Rigidity to Recurrency (CVPR 2018 · similarity 0.88)
- PR-RRN: Pairwise-Regularized Residual-Recursive Networks for Non-Rigid Structure-From-Motion (ICCV 2021 · similarity 0.88)
- Neural Dense Non-Rigid Structure from Motion with Latent Space Constraints (ECCV 2020 · similarity 0.88)
- Reconstruct as Far as You Can: Consensus of Non-Rigid Reconstruction from Feasible Regions (TPAMI 2019 · similarity 0.88)
- Deep Non-Rigid Structure-From-Motion: A Sequence-to-Sequence Translation Perspective (TPAMI 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.
