Research map / 3D vision / 3D generation
3D Scene Assertion Verification
Jun Lin, Jiayu Ding, Xiangtian Si, Xitong Cao, Lixin Hong, Zhang Chen, Chenxi Lv, Wenqian Wang
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 Early LPSS routing (ablation) | backbone · 3DSAV test set | accuracy: 77.3 -> 81.9 | Improved |
| add Late LPSS routing (ablation) | backbone · 3DSAV test set | accuracy: 77.3 -> 82.0 | Improved |
| combine both stages (full DualLPSS) (ablation) | single-stage variants · 3DSAV test set | accuracy: 82.0 -> 82.5 | Improved |
| add intra-text structural strength (ablation) | DualLPSS · 3DSAV test set | accuracy: 81.7 -> 82.5 | Improved |
| add intra-object structural strength (ablation) | DualLPSS · 3DSAV test set | accuracy: 82.2 -> 82.5 | Improved |
| use Gumbel-Softmax instead of standard Softmax in routing (ablation) | DualLPSS · 3DSAV test set | accuracy: 81.7 -> 82.5 | Improved |
| use Soft Top-2 instead of Hard Top-K routing (ablation) | DualLPSS · 3DSAV test set | accuracy: 81.2 -> 82.5 | Improved |
| set subspace count K=8 instead of 6 (ablation) | Early LPSS · 3DSAV test set | accuracy: -1.3% | Hurt |
| set subspace count K=8 instead of 6 (ablation) | Late LPSS · 3DSAV test set | accuracy: -1.0% | Hurt |
| use DualLPSS instead of Chat-Scene | 3DSAV baselines · 3DSAV test set | accuracy: 80.5 -> 82.5 | Improved |
| use DualLPSS instead of Chat-Scene on converted ScanQA | cross-dataset transfer · 3DSAV test set | accuracy: 70.9 -> 82.1 | Improved |
| use DualLPSS instead of Chat-Scene on Quantity assertions | 3DSAV per-type · 3DSAV test set | Quantity accuracy: 95.3 -> 90.2 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ScanQA — ScanQA: 3D Question Answering for Spatial Scene Understanding (CVPR 2022) · better on 2 of 2 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
- Bridging the Gap between 2D and 3D Visual Question Answering: A Fusion Approach for 3D VQA (AAAI 2024 · similarity 0.91)
- DSPNet: Dual-vision Scene Perception for Robust 3D Question Answering (CVPR 2025 · similarity 0.89)
- 3D Question Answering via only 2D Vision-Language Models (ICML 2025 · similarity 0.89)
- Hypo3D: Exploring Hypothetical Reasoning in 3D (ICML 2025 · similarity 0.89)
- Unveiling the Mist over 3D Vision-Language Understanding: Object-centric Evaluation with Chain-of-Analysis (CVPR 2025 · similarity 0.89)
- MMScan: A Multi-Modal 3D Scene Dataset with Hierarchical Grounded Language Annotations (NeurIPS 2024 · 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.
