Research map / Detection and segmentation / 3D detection and driving
Collaborative Novel Object Discovery and Box-Guided Cross-Modal Alignment for Open-Vocabulary 3D Object Detection
Yang Cao, Yihan Zeng, Hang Xu, Dan Xu
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 3D novel object discovery (3D-NOD) (ablation) | Distillation base · SUN-RGBD | AP_Novel: 3.28 -> 5.48 | Improved |
| add 3D-NOD (ablation) | Distillation base · SUN-RGBD | AR_Novel: 19.87 -> 33.45 | Improved |
| add discovery-driven alignment (DCMA) instead of plain alignment (ablation) | 3D-NOD + Distillation · SUN-RGBD | AP_Novel: 0.85 -> 6.71 | Improved |
| use class-agnostic distillation instead of class-specific contrastive learning (ablation) | DCMA · SUN-RGBD | AP_Novel: 5.35 -> 6.71 | Improved |
| use class-specific contrastive learning instead of class-agnostic distillation (ablation) | DCMA · SUN-RGBD | AR_Novel: 33.66 -> 37.93 | Hurt |
| use bipartite matching instead of 3DIoU matching (ablation) | DCMA · SUN-RGBD | AP_Novel: 5.68 -> 6.71 | Improved |
| add DCMA to 3D-CLIP + 3D-NOD (ablation) | 3D-NOD · SUN-RGBD | AP_Base: 26.08 -> 38.72 | Improved |
| add Enrichment (3D-NODE, 5 inserted objects) (ablation) | 3D-NOD + DCMA · SUN-RGBD | AP_Novel: 6.71 -> 7.53 | Improved |
| insert 10 instead of 5 novel objects in Enrichment (ablation) | 3D-NODE + DCMA · SUN-RGBD | AP_Novel: 7.53 -> 7.29 | Hurt |
| add box guidance (Box-DCMA) (ablation) | 3D-NODE + DCMA · SUN-RGBD | AP_Novel: 7.53 -> 9.17 | Improved |
| match background boxes with CLIP instead of box guidance (ablation) | 3D-NODE + DCMA · SUN-RGBD | AP_Novel: 7.53 -> 6.47 | Hurt |
| set 3D-NOD thresholds to 0.3/0.3 instead of disabling priors (ablation) | 3D-NOD + DCMA · SUN-RGBD | AP_Novel: 3.28 -> 6.71 | Improved |
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.
- Det-CLIP2 — DetCLIP: Dictionary-Enriched Visual-Concept Paralleled Pre-training for Open-world Detection (NeurIPS 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
- Hierarchical Cross-Modal Alignment for Open-Vocabulary 3D Object Detection (AAAI 2025 · similarity 0.91)
- ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · similarity 0.90)
- Towards 3D Objectness Learning in an Open World (NeurIPS 2025 · similarity 0.89)
- FM-OV3D: Foundation Model-Based Cross-Modal Knowledge Blending for Open-Vocabulary 3D Detection (AAAI 2024 · similarity 0.88)
- Training an Open-Vocabulary Monocular 3D Detection Model without 3D Data (NeurIPS 2024 · similarity 0.88)
- LabelAny3D: Label Any Object 3D in the Wild (NeurIPS 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.
