Research map / Detection and segmentation / Open-set and few-shot detection
Beyond Reassembly: Fractured Object Recovery with Missing Parts
Qun-Ce Xu, Jiahui Li, Yan-Pei Cao, Weihao Cheng, Tai-Jiang Mu, Ying Shan, Chuan Li, Da Chen, Yong-Liang Yang, Shi-min Hu
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 |
|---|---|---|---|
| jointly learn pose estimation and missing-part shape prediction instead of two stages (ablation) | Ours · Fractured object recovery | Emissing (x10^3): 101.58 -> 39.11 | Improved |
| jointly learn pose estimation and missing-part shape prediction instead of two stages (ablation) | Ours · Fractured object recovery | Erot (x10^3): 75.17 -> 18.71 | Improved |
| add rotation-invariant mapping to the input point clouds (ablation) | Ours · Fractured object recovery | Erot (x10^3): 27.66 -> 18.71 | Improved |
| add consistency loss between predicted existing-part shapes and aligned parts (ablation) | Ours · Fractured object recovery | Emissing (x10^3): 59.43 -> 39.11 | Improved |
| add consistency loss between predicted existing-part shapes and aligned parts (ablation) | Ours · Fractured object recovery | Erot (x10^3): 22.83 -> 18.71 | Improved |
| use point cloud shape representation instead of SDF implicit surfaces for missing parts (ablation) | Ours · Fractured object recovery | Emissing (x10^3): 87.64 -> 39.11 | Improved |
| use joint model instead of best pipeline baseline GARF w/ AdaPoinTr | fractured object recovery · Fractured object recovery | Emissing (x10^3): 46.27 -> 39.11 | Improved |
Closest papers by meaning
- Jigsaw++: Imagining Complete Shape Priors for Object Reassembly (ICCV 2025 · similarity 0.89)
- Image Reassembly Combining Deep Learning and Shortest Path Problem (ECCV 2018 · similarity 0.88)
- Learning by Restoring Broken 3D Geometry (TPAMI 2023 · similarity 0.88)
- DiffAssemble: A Unified Graph-Diffusion Model for 2D and 3D Reassembly (CVPR 2024 · similarity 0.88)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.87)
- PHFormer: Multi-Fragment Assembly Using Proxy-Level Hybrid Transformer (AAAI 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.
