Research map / Detection and segmentation / Open-set and few-shot detection
3D Open-Vocabulary Panoptic Segmentation with 2D-3D Vision-Language Distillation
Zihao Xiao, Longlong Jing, Shangxuan Wu, Alex Zihao Zhu, Jingwei Ji, Chiyu Max Jiang, Wei-Chih Hung, Thomas Funkhouser, Weicheng Kuo, Anelia Angelova, Yin Zhou, Shiwei Sheng
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 voxel-level distillation loss L_V (ablation) | Ours · nuScenes val (B12/N4) | PQ: 56.3 -> 62.0 | Improved |
| add voxel-level distillation loss L_V (ablation) | Ours · nuScenes val (B12/N4) | PQ novel stuff: 0.2 -> 35.2 | Improved |
| add object-level distillation loss L_O (ablation) | PFC + QA + fusion · nuScenes val (B12/N4) | PQ novel things: 38.1 -> 43.8 | Improved |
| add object-level distillation loss L_O (ablation) | PFC + QA + fusion · nuScenes val (B12/N4) | PQ: 56.4 -> 56.3 | No clear effect |
| add multimodal LiDAR + CLIP feature fusion (ablation) | PFC + QA · nuScenes val (B12/N4) | PQ: 55.5 -> 56.4 | Improved |
| add fixed query set for base stuff classes (query assignment) (ablation) | PFC · nuScenes val (B12/N4) | PQ: 54.8 -> 55.5 | Improved |
| add query assignment (effect on novel classes) (ablation) | PFC · nuScenes val (B12/N4) | PQ novel things: 37.3 -> 35.7 | Hurt |
| replace 2D-style open-vocab model PFC with proposed fusion, single head and distillation losses | PFC baseline · SemanticKITTI val (B14/N5) | PQ: 33.7 -> 42.2 | Improved |
| use a different base/novel split with 5 novel classes | Ours · nuScenes val (B11/N5) | PQ: 62.0 -> 52.8 | Hurt |
| train P3Former with base classes only vs open-vocab method | P3Former · nuScenes val base stuff | PQ base stuff: 64.2 vs 69.2 (Ours) | Hurt |
Closest papers by meaning
- OpenMask3D: Open-Vocabulary 3D Instance Segmentation (NeurIPS 2023 · similarity 0.91)
- 3D-AVS: LiDAR-based 3D Auto-Vocabulary Segmentation (CVPR 2025 · similarity 0.90)
- PanoGS: Gaussian-based Panoptic Segmentation for 3D Open Vocabulary Scene Understanding (CVPR 2025 · similarity 0.90)
- Segment Any 3D Object with Language (ICLR 2025 · similarity 0.90)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.89)
- Mosaic3D: Foundation Dataset and Model for Open-Vocabulary 3D Segmentation (CVPR 2025 · similarity 0.88)
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.
