Research map / Detection and segmentation / Open-set and few-shot detection
XMask3D++: Cross-Modal Mask Reasoning for Open Vocabulary 3D Segmentation
Ziyi Wang, Yanbo Wang, Xumin Yu, Zheng Dong Zhu, Jie Zhou, Jiwen Lu
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 fine-grained mask-level regularization loss L_mask (ablation) | XMask3D 3D branch · ScanNet B12/N7 | novel mIoU: +11.7 | Improved |
| add fine-grained mask-level regularization loss L_mask (ablation) | XMask3D fused output · ScanNet B12/N7 | novel mIoU: +1.5 | Improved |
| use implicit 3D condition for the diffusion mask generator instead of text or implicit 2D condition (ablation) | XMask3D · ScanNet B12/N7 | novel mIoU | Improved |
| fuse 2D and 3D branch features instead of using either branch alone (ablation) | XMask3D · ScanNet B12/N7 | base and novel mIoU | Improved |
| use XMask3D instead of PLA | XMask3D · ScanNet200 B170/N30 | novel mIoU: 7.8 -> 13.3 | Improved |
| use XMask3D instead of PLA | XMask3D · ScanNet B12/N7 | hIoU: 55.3 -> 61.7 | Improved |
| use XMask3D instead of PLA | XMask3D · S3DIS B8/N4 | novel mIoU: 24.5 -> 37.2 | Improved |
| use XMask3D instead of PLA | XMask3D · S3DIS B6/N6 | base mIoU: 55.5 -> 52.8 | Hurt |
| use XMask3D instead of OV3D | XMask3D · ScanNet B15/N4 | hIoU: 72.4 -> 70.0 | Hurt |
| use XMask3D instead of OpenScene | XMask3D · ScanNet200 B150/N50 | hIoU: 15.2 -> 15.5 | No clear effect |
Closest papers by meaning
- XMask3D: Cross-modal Mask Reasoning for Open Vocabulary 3D Semantic Segmentation (NeurIPS 2024 · similarity 0.99)
- Segment Any 3D Object with Language (ICLR 2025 · similarity 0.91)
- Mosaic3D: Foundation Dataset and Model for Open-Vocabulary 3D Segmentation (CVPR 2025 · similarity 0.90)
- Open-Vocabulary 3D Semantic Segmentation with Foundation Models (CVPR 2024 · similarity 0.89)
- MaskClustering: View Consensus based Mask Graph Clustering for Open-Vocabulary 3D Instance Segmentation (CVPR 2024 · similarity 0.88)
- Masked Point-Entity Contrast for Open-Vocabulary 3D Scene Understanding (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.
