Research map / 3D vision / 3D generation
LR-AdaInSeg:Adaptive Instance Segmentation of Incomplete 3D Scenes Driven by Low-Rank Networks
Qin Zhang, Kun Zhou, Xulun Ye
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 Bayesian Low-Rank Module (ablation) | LR-AdaInSeg · ScanNetv2 (lambda=0.8) | AP50: 43.2 -> 48.6 | Improved |
| add Bayesian Low-Rank Module (ablation) | LR-AdaInSeg · ScanNet++ (lambda=0.8) | AP50: 27.4 -> 34.1 | Improved |
| add CGCS contrastive graph class similarity loss (ablation) | LR-AdaInSeg · ScanNetv2 (lambda=0.8) | AP50: 45.7 -> 48.6 | Improved |
| add CGCS loss (ablation) | LR-AdaInSeg · ScanNet++ (lambda=0.8) | AP50: 31.5 -> 34.1 | Improved |
| add Bayesian Low-Rank Module (ablation) | LR-AdaInSeg · ScanNet200 (lambda=1) | AP: 42.8 -> 48.6 | Improved |
| LR-AdaInSeg vs SAM2Object on incomplete ScanNet++ | LR-AdaInSeg · ScanNet++ (incomplete scenes, lambda=0.7) | AP50: 30.7 -> 33.5 | Improved |
| LR-AdaInSeg vs SAM2Object on incomplete ScanNetv2 | LR-AdaInSeg · ScanNetv2 (incomplete scenes, lambda=0.7) | AP: 25.1 -> 23.7 | Hurt |
| LR-AdaInSeg vs SAM2Object on complete ScanNetv2 | LR-AdaInSeg · ScanNetv2 (complete scenes) | AP: 34.0 -> 28.7 | Hurt |
| LR-AdaInSeg vs SAMPro3D on complete ScanNet200 | LR-AdaInSeg · ScanNet200 (complete scenes) | AP: 26.3 -> 24.7 | Hurt |
| increase scene completeness lambda (ablation) | LR-AdaInSeg · ScanNet series | AP | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 on most shared results.
- SAM2Object — SAM2Object: Consolidating View Consistency via SAM2 for Zero-Shot 3D Instance Segmentation (CVPR 2025) · better on 1 of 4 shared results
- Segment3D — Segment3D: Learning Fine-Grained Class-Agnostic 3D Segmentation without Manual Labels (ECCV 2024) · better on 0 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding (AAAI 2025 · similarity 0.85)
- No Time to Train: Empowering Non-Parametric Networks for Few-shot 3D Scene Segmentation (CVPR 2024 · similarity 0.85)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.85)
- SegGraph: Leveraging Graphs of SAM Segments for Few-Shot 3D Part Segmentation (NeurIPS 2025 · similarity 0.85)
- SAI3D: Segment Any Instance in 3D Scenes (CVPR 2024 · similarity 0.84)
- OnlineAnySeg: Online Zero-Shot 3D Segmentation by Visual Foundation Model Guided 2D Mask Merging (CVPR 2025 · similarity 0.84)
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.
