Research map / 3D vision / Face and shape
K-Best Transformation Synchronization
Yifan Sun, Jiacheng Zhuo, Arnav Mohan, Qixing Huang
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 |
|---|---|---|---|
| use K-best propagation and clustering instead of MRF-SFM | transformation synchronization · ShapeNetCoreSym | mean rotation error: 12.1 -> 6.7 | Improved |
| use K-best propagation and clustering instead of MRF-SFM | transformation synchronization · ScanNetSym | mean rotation error: 31.6 -> 19.6 | Improved |
| use K-best propagation and clustering instead of best baseline | transformation synchronization · ShapeNetCore (single-best) | mean rotation error: 13.7 -> 9.7 | Improved |
| use K-best propagation and clustering instead of best baseline | transformation synchronization · ScanNet (single-best) | mean rotation error: 36.3 -> 21.3 | Improved |
| use ground-truth K instead of predicted K (ablation) | K-best sync · ShapeNetCoreSym | mean rotation error: 6.7 -> 5.8 | Improved |
| use ground-truth K instead of predicted K (ablation) | K-best sync · ShapeNetCore (single-best) | mean rotation error: 9.7 -> 15.2 | Hurt |
| use 3x coarser sample bin size (ablation) | K-best sync · ShapeNetCoreSym | mean rotation error: 6.7 -> 8.9 | Hurt |
| use 0.5x finer sample bin size (ablation) | K-best sync · ShapeNetCoreSym | mean rotation error: 6.7 -> 6.5 | No clear effect |
| vary the root object (ablation) | K-best sync · all benchmarks | mean rotation error | No clear effect |
| use ours instead of RobustAlign in speed | K-best sync · ShapeNetCore | running time: 0.6s -> 1.2s | Hurt |
| use ours instead of RobustAlign in speed | K-best sync · ScanNet | running time: 0.8s -> 4.6s | Hurt |
Closest papers by meaning
- Learning Transformation Synchronization (CVPR 2019 · similarity 0.88)
- Joint Map and Symmetry Synchronization (ECCV 2018 · similarity 0.86)
- Probabilistic Permutation Synchronization Using the Riemannian Structure of the Birkhoff Polytope (CVPR 2019 · similarity 0.86)
- Progressive Correspondence Regenerator for Robust 3D Registration (CVPR 2025 · similarity 0.84)
- Tensor-Based Synchronization and the Low-Rankness of the Block Trifocal Tensor (NeurIPS 2024 · similarity 0.83)
- Multi-Shape Matching with Cycle Consistency Basis via Functional Maps (AAAI 2025 · similarity 0.83)
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.
