Research map / Optimization / Nonconvex and smooth optimization
Accurate Optimization of Weighted Nuclear Norm for Non-Rigid Structure from Motion
José Pedro Iglesias, Carl Olsson, Marcus Valtonen Örnhag
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 |
|---|---|---|---|
| refine ADMM solution with bilinear parameterization and Levenberg-Marquardt (weighted nuclear norm) | ADMM-WNN · NRSfM Challenge Balloon (perspective) | average log-loss: -2.617 -> -2.931 | Improved |
| refine ADMM solution with bilinear parameterization (nuclear norm) | ADMM-NN · NRSfM Challenge Tearing (perspective) | average log-loss: -1.471 -> -2.053 | Improved |
| refine ADMM solution with bilinear parameterization (weighted nuclear norm) | ADMM-WNN · NRSfM Challenge Paper (perspective) | 3D reconstruction error (mm): 10.14 -> 5.68 | Improved |
| refine ADMM solution with bilinear parameterization (nuclear norm) | ADMM-NN · NRSfM Challenge Balloon (perspective) | 3D reconstruction error (mm): 8.05 -> 6.48 | Improved |
| refine ADMM solution with bilinear parameterization (nuclear norm) | ADMM-NN · NRSfM Challenge Stretch (perspective) | 3D reconstruction error (mm): 6.04 -> 6.00 | No clear effect |
| use bilinear refinement with nuclear norm instead of BMM on orthogonal data | BMM · NRSfM Challenge Paper (orthogonal) | 3D reconstruction error (mm): 8.94 -> 10.94 | Hurt |
| use bilinear refinement with nuclear norm instead of BMM on orthogonal data | BMM · NRSfM Challenge Paper (orthogonal) | average log-loss: -1.712 -> -2.188 | Improved |
| use weighted nuclear norm in place of nuclear norm with bilinear refinement (ablation) | Ours · NRSfM Challenge Paper (perspective) | 3D reconstruction error (mm): 6.80 -> 5.68 | Improved |
| use weighted nuclear norm in place of nuclear norm with bilinear refinement (ablation) | Ours · NRSfM Challenge Articulated (perspective) | 3D reconstruction error (mm): 16.13 -> 16.53 | Hurt |
| use ADMM-only weighted nuclear norm instead of nuclear norm (ablation) | ADMM · NRSfM Challenge Paper (perspective) | 3D reconstruction error (mm): 7.96 -> 10.14 | Hurt |
| use weighted nuclear norm with bilinear refinement on orthogonal data | R-BMM · NRSfM Challenge Articulated (orthogonal) | 3D reconstruction error (mm): 16.00 -> 15.03 | Improved |
| refine ADMM with bilinear parameterization on Tearing-Zigzag sequence | ADMM-WNN · NRSfM Challenge Tearing Zigzag (perspective) | 3D reconstruction error (mm): 7.09 -> 8.12 | Hurt |
Closest papers by meaning
- A Unified Optimization Framework for Low-Rank Inducing Penalties (CVPR 2020 · similarity 0.89)
- Bilinear Parameterization for Non-Separable Singular Value Penalties (CVPR 2021 · similarity 0.87)
- Non-Convex Rank/Sparsity Regularization and Local Minima (ICCV 2017 · similarity 0.85)
- Near-Optimal Weighted Matrix Completion (JMLR 2023 · similarity 0.85)
- Inertial Block Proximal Methods for Non-Convex Non-Smooth Optimization (ICML 2020 · similarity 0.85)
- Cauchy-Schwarz Regularizers (ICLR 2025 · similarity 0.82)
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.
