Research map / 3D vision / Depth and stereo
Self-Expressive Dictionary Learning for Dynamic 3D Reconstruction
Enliang Zheng, Dinghuang Ji, Enrique Dunn, Jan‐Michael Frahm
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 |
|---|---|---|---|
| lower per-camera frame rate (downsample from 120 Hz to 30 Hz total) | Self-expressive dictionary learning · Synthetic mocap | reconstruction accuracy: more than 97% of 3D points still have errors less than 5 cm at 7.5 Hz | Hurt |
| allow temporally consecutive captures to be assigned to the same camera (unconstrained assignment) | Self-expressive dictionary learning · Synthetic mocap | reconstruction accuracy | Hurt |
| add Gaussian noise to 2D measurements | Self-expressive dictionary learning (soft constraint, lambda3=100) · Synthetic mocap | reconstruction accuracy | Hurt |
| use soft instead of hard viewing-ray constraint under noise (ablation) | Self-expressive dictionary learning · Synthetic mocap, noisy measurements | reconstruction accuracy | No clear effect |
| increase missing-data rate to 40% | Self-expressive dictionary learning · Synthetic mocap | % of points with error < 30 mm: 94.38% of points having errors less than 30 mm at 40% occlusion | Mixed |
| compare against nuclear-norm NRSFM (Dai et al.) adapted to perspective cameras | Self-expressive dictionary learning · Synthetic mocap, random camera configuration, no noise | qualitative reconstruction | Improved |
| compare against trajectory triangulation with one handheld camera (large system condition) | Self-expressive dictionary learning · Synthetic mocap, handheld camera | system condition / error: 1/sigma_min=2228 | Improved |
Closest papers by meaning
- Prior-Less Compressible Structure From Motion (CVPR 2016 · similarity 0.89)
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (NeurIPS 2024 · similarity 0.85)
- Sparse3D: Distilling Multiview-Consistent Diffusion for Object Reconstruction from Sparse Views (AAAI 2024 · similarity 0.85)
- Novel View Synthesis from A Few Glimpses via Test-Time Natural Video Completion (NeurIPS 2025 · similarity 0.85)
- ReMatching Dynamic Reconstruction Flow (ICLR 2025 · similarity 0.85)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 2024 · similarity 0.85)
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.
