Research map / Continual learning / Class-incremental
Seeing 3D Through 2D Lenses: 3D Few-Shot Class-Incremental Learning via Cross-Modal Geometric Rectification
Tuo Xiang, Xuemiao Xu, Bangzhen Liu, Jinyi Li, Yong Li, Shengfeng He
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 SAGR to the baseline (ablation) | CMGR baseline · ShapeNet->CO3D | AA: 72.3 -> 74.3 | Improved |
| add TAM to the baseline (ablation) | CMGR baseline · ShapeNet->CO3D | AA: 72.3 -> 72.9 | Improved |
| add BND to the baseline (ablation) | CMGR baseline · ShapeNet->CO3D | AA: 72.3 -> 74.3 | Improved |
| add SAGR to TAM+BND (ablation) | CMGR (TAM+BND) · ShapeNet->CO3D | AA: 75.1 -> 78.6 | Improved |
| add TAM to SAGR+BND (ablation) | CMGR (SAGR+BND) · ShapeNet->CO3D | AA: 76.4 -> 78.6 | Improved |
| add BND to SAGR+TAM (ablation) | CMGR (SAGR+TAM) · ShapeNet->CO3D | AA: 75.3 -> 78.6 | Improved |
| use all three modules versus baseline (ablation) | CMGR · ShapeNet->CO3D | Delta_A: 5.4 -> 3.2 | Improved |
| CMGR versus best prior method C3PR | CMGR · ShapeNet->CO3D | AA: 72.0 -> 78.6 | Improved |
| CMGR versus FILP-3D on final-task accuracy | CMGR · ShapeNet->CO3D | final-task accuracy: 57.1 -> 66.3 | Improved |
| CMGR versus C3PR on forgetting rate | CMGR · ShapeNet->CO3D | Delta_A: 2.8 -> 3.2 | Hurt |
| CMGR versus C3PR on forgetting rate | CMGR · ModelNet->ModelNet | Delta_A: 6.2 -> 8.7 | Hurt |
| CMGR versus Microshape on within-domain ModelNet | CMGR · ModelNet->ModelNet | AA: 79.6 -> 79.8 | No clear effect |
Closest papers by meaning
- Learning 3D Representations for Spatial Intelligence from Unposed Multi-View Images (CVPR 2026 · similarity 0.86)
- TAMM: TriAdapter Multi-Modal Learning for 3D Shape Understanding (CVPR 2024 · similarity 0.86)
- 3DCoMPaT200: Language Grounded Large-Scale 3D Vision Dataset for Compositional Recognition (NeurIPS 2024 · similarity 0.85)
- CrossOver: 3D Scene Cross-Modal Alignment (CVPR 2025 · similarity 0.85)
- MM-Mixing: Multi-Modal Mixing Alignment for 3D Understanding (AAAI 2025 · similarity 0.85)
- CLIP-Driven Open-Vocabulary 3D Scene Graph Generation via Cross-Modality Contrastive Learning (CVPR 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.
