Research map / Point clouds / Point cloud classification
InstanceRefer: Cooperative Holistic Understanding for Visual Grounding on Point Clouds Through Instance Multi-Level Contextual Referring
Zhihao Yuan, Xu Yan, Yinghong Liao, Ruimao Zhang, Sheng Wang, Zhen Li, Shuguang Cui
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 the RP module to the AP module (ablation) | InstanceRefer · ScanRefer val | Overall Acc@0.5: 29.46 -> 30.49 | Improved |
| add the GLP module to AP+RP (ablation) | InstanceRefer · ScanRefer val | Overall Acc@0.5: 30.49 -> 31.04 | Improved |
| add the co-attention matching module over cosine-similarity ranking (ablation) | InstanceRefer · ScanRefer val | Overall Acc@0.5: 31.04 -> 32.93 | Improved |
| use RP instead of GLP as the single module (ablation) | InstanceRefer · ScanRefer val | Overall Acc@0.5: 25.29 -> 28.01 | Improved |
| use AP instead of RP as the single module (ablation) | InstanceRefer · ScanRefer val | Overall Acc@0.5: 28.01 -> 29.46 | Improved |
| replace ScanRefer proposals with PointGroup panoptic instances | ScanRefer · ScanRefer val | Overall Acc@0.5: 24.32 -> 29.07 | Improved |
| use instance point clouds instead of points in GT boxes | ScanRefer · ScanRefer val (GT) | Overall Acc@0.5: 40.06 -> 44.48 | Improved |
| use GT language classification instead of predicted (ablation) | InstanceRefer · ScanRefer val | Overall Acc@0.5: 32.93 -> 33.35 | No clear effect |
| replace GRU with pre-trained BERT language encoder | TGNN · ScanRefer val | Overall Acc@0.5: 27.92 -> 29.70 | Improved |
| swap ScanRefer/ReferIt3DNet for InstanceRefer on PG instances | PG + ScanRefer · ScanRefer val | Overall Acc@0.5: 29.07 -> 32.93 | Improved |
Later papers that compared with this method
3 later accepted papers report InstanceRefer as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- EDA: Explicit Text-Decoupling and Dense Alignment for 3D Visual Grounding (CVPR 2023 · better on 2 of 2 shared results)
- Free-Form Description Guided 3D Visual Graph Network for Object Grounding in Point Cloud (ICCV 2021 · better on 7 of 7 shared results)
- 3DVG-Transformer: Relation Modeling for Visual Grounding on Point Clouds (ICCV 2021 · better on 3 of 4 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
- Four Ways to Improve Verbo-visual Fusion for Dense 3D Visual Grounding (ECCV 2024 · similarity 0.92)
- From Objects to Anywhere: A Holistic Benchmark for Multi-level Visual Grounding in 3D Scenes (NeurIPS 2025 · similarity 0.91)
- CoT3DRef: Chain-of-Thoughts Data-Efficient 3D Visual Grounding (ICLR 2024 · similarity 0.89)
- AugRefer: Advancing 3D Visual Grounding via Cross-Modal Augmentation and Spatial Relation-based Referring (AAAI 2025 · similarity 0.89)
- Grounded 3D-Aware Spatial Vision-Language Modeling (CVPR 2026 · similarity 0.89)
- ReasonGrounder: LVLM-Guided Hierarchical Feature Splatting for Open-Vocabulary 3D Visual Grounding and Reasoning (CVPR 2025 · similarity 0.89)
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.
