Research map / Detection and segmentation / Open-set and few-shot detection
Lowis3D: Language-Driven Open-World Instance-Level 3D Scene Understanding
Runyu Ding, Jihan Yang, Chuhui Xue, Wenqing Zhang, Song Bai, Xiaojuan Qi
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 binary calibration module (ablation) | OV-SparseConvNet · ScanNet B15/N4 semantic seg | hIoU: 00.0 -> 39.8 | Improved |
| add scene-level, view-level or entity-level point-language supervision to binary calibration (ablation) | Lowis3D · ScanNet semantic seg | hIoU: 39.8 -> 54.6 / 61.3 / 63.6 | Improved |
| add entity-level instead of scene-level caption supervision (ablation) | Lowis3D · ScanNet B13/N4 instance seg | hAP50: 52.8 -> 56.6 | Improved |
| add debiased instance localization (DIL) (ablation) | Lowis3D · ScanNet B13/N4 instance seg | hAP50: 54.9 -> 59.1 | Improved |
| add DIL (ablation) | Lowis3D · ScanNet B13/N4 offset error | mAE novel: 0.69 -> 0.48 | Improved |
| add DIL (ablation) | Lowis3D · ScanNet B13/N4 proposals | hAR: 44.7 -> 47.9 | Improved |
| add language supervision (PLA w/o Cap. vs PLA / Lowis3D) (ablation) | PLA · ScanNet B15/N4 semantic seg | hIoU: 39.7 -> 65.3 | Improved |
| use CLIP text encoder instead of BERT (ablation) | Lowis3D · ScanNet B15/N4 semantic seg | hIoU: 61.2 -> 65.3 | Improved |
| use CLIP text encoder instead of GPT2 (ablation) | Lowis3D · ScanNet B15/N4 semantic seg | hIoU: 61.0 -> 65.3 | Improved |
| use OFA instead of ViT-GPT2 captioner (ablation) | Lowis3D · ScanNet B10/N9 semantic seg | hIoU: 53.1 -> 56.6 | Improved |
| keep only label names in captions instead of full caption (ablation) | Lowis3D · ScanNet B15/N4 semantic seg | hIoU: 65.3 -> 57.6 | Hurt |
| keep only entities instead of full caption (ablation) | Lowis3D · ScanNet B15/N4 semantic seg | hIoU: 65.3 -> 65.7 | No clear effect |
Later papers that compared with this method
1 later accepted paper reports Lowis3D as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · better on 1 of 1 shared result)
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
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.92)
- OpenLex3D: A Tiered Benchmark for Open-Vocabulary 3D Scene Representations (NeurIPS 2025 · similarity 0.92)
- Segment Any 3D Object with Language (ICLR 2025 · similarity 0.91)
- RegionPLC: Regional Point-Language Contrastive Learning for Open-World 3D Scene Understanding (CVPR 2024 · similarity 0.91)
- Cross-Modal and Uncertainty-Aware Agglomeration for Open-Vocabulary 3D Scene Understanding (CVPR 2025 · similarity 0.90)
- Mosaic3D: Foundation Dataset and Model for Open-Vocabulary 3D Segmentation (CVPR 2025 · similarity 0.90)
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.
