Research map / 3D vision / Depth and stereo
Structure-From-Sherds: Incremental 3D Reassembly of Axially Symmetric Pots From Unordered and Mixed Fragment Collections
Je Hyeong Hong, Seong Jong Yoo, Muhammad Arshad Zeeshan, Young Min Kim, Jinwook Kim
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 beam search (ablation) | Structure-from-Sherds · A+B+C multi-pot reassembly (ablation) | sherds correctly reassembled: 11/21 -> 19/21 | Improved |
| add rim constraint (ablation) | Structure-from-Sherds · A+B+C multi-pot reassembly (ablation) | sherds correctly reassembled: 18/21 -> 19/21 | Improved |
| add thickness constraint (ablation) | Structure-from-Sherds · A+B+C multi-pot reassembly (ablation) | sherds correctly reassembled: 18/21 -> 19/21 | Improved |
| add rim constraint to prune initial matches (ablation) | Structure-from-Sherds · A+B+C multi-pot reassembly (ablation) | # initial pairwise matches: 634 -> 483 | Improved |
| add thickness constraint to prune initial matches (ablation) | Structure-from-Sherds · A+B+C multi-pot reassembly (ablation) | # initial pairwise matches: 943 -> 483 | Improved |
| increase beam search from b=3,k=5 to b=10,k=20 (ablation) | Structure-from-Sherds · Pot sherd dataset, pot All (80) | sherds correctly reassembled: 43 / 80 -> 66 / 80 | Improved |
| increase beam from b=3,k=5 to b=5,k=10 (ablation) | Structure-from-Sherds · Pot sherd dataset, pot D+E (59) | sherds correctly reassembled: 27 / 59 -> 36 / 59 | Improved |
| increase beam from b=5,k=10 to b=10,k=20 (ablation) | Structure-from-Sherds · Pot sherd dataset, pot E (31) | sherds correctly reassembled: 26 / 31 -> 24 / 31 | Hurt |
| increase beam size on single pots (ablation) | Structure-from-Sherds · Pot sherd dataset, pots A, B, C | sherds correctly reassembled: 8 / 8 -> 8 / 8 | No clear effect |
| increase beam from b=3,k=5 to b=10,k=20 (ablation) | Structure-from-Sherds · Pot sherd dataset, pot All (80) | preprocessing runtime (min.): 167.5 -> 592.4 | Hurt |
| compare to Son et al. method | Structure-from-Sherds · Pot sherd dataset (different data) | reassembly accuracy: 31/48 -> over 17.9% improvement | Improved |
Closest papers by meaning
- Scalable Geometric Fracture Assembly via Co-creation Space among Assemblers (AAAI 2024 · similarity 0.85)
- ColabSfM: Collaborative Structure-from-Motion by Point Cloud Registration (CVPR 2025 · similarity 0.85)
- PuzzleFusion++: Auto-agglomerative 3D Fracture Assembly by Denoise and Verify (ICLR 2025 · similarity 0.84)
- Leveraging Camera Triplets for Efficient and Accurate Structure-from-Motion (CVPR 2024 · similarity 0.84)
- Light3R-SfM: Towards Feed-forward Structure-from-Motion (CVPR 2025 · similarity 0.83)
- SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation (NeurIPS 2026 · 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.
