Research map / Point clouds / Segmentation and detection
3D Spatial Recognition Without Spatially Labeled 3D
Zhongzheng Ren, Ishan Misra, Alexander G. Schwing, Rohit Girdhar
Check 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 |
|---|---|---|---|
| Jointly train segmentation and detection instead of using single-task MIL-seg | WyPR · ScanNet val | mIoU: 20.7 -> 29.6 | Improved |
| Add external shape and floor-height priors | WyPR · ScanNet val | mIoU: 29.6 -> 31.1 | Improved |
| Add external object prior | WyPR · ScanNet | mAP: +1.4% | Improved |
| Add self-training losses to segmentation and detection (ablation) | WyPR · ScanNet val | mIoU: 22.1 -> 29.6 | Improved |
| Add self-training losses to segmentation and detection (ablation) | WyPR · ScanNet val | mAP: 13.2 -> 18.3 | Improved |
| Add cross-transformation consistency losses (ablation) | WyPR · ScanNet val | mAP: 16.9 -> 18.3 | Improved |
| Add detection-to-segmentation cross-task consistency (ablation) | WyPR · ScanNet val | mIoU: 26.7 -> 29.6 | Improved |
| Add local smoothness over primitive structures (ablation) | WyPR · ScanNet val | mIoU: 27.3 -> 29.6 | Improved |
| Add local smoothness over primitive structures (ablation) | WyPR · ScanNet val | mAP: 17.8 -> 18.3 | Improved |
| Add segmentation similarity to GSS proposal grouping (ablation) | GSS · ScanNet val | AR: 86.2 -> 89.3 | Improved |
| Add segmentation similarity to GSS proposal grouping (ablation) | GSS · ScanNet val | MABO: 0.378 -> 0.409 | Improved |
| Jitter points before GSS proposal generation (ablation) | GSS · ScanNet | AP: >2% | Improved |
Closest papers by meaning
- A Simple Vision Transformer for Weakly Semi-supervised 3D Object Detection (ICCV 2023 · similarity 0.91)
- Leveraging Imagery Data with Spatial Point Prior for Weakly Semi-supervised 3D Object Detection (AAAI 2024 · similarity 0.90)
- WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection (AAAI 2024 · similarity 0.89)
- Multi-Modality Affinity Inference for Weakly Supervised 3D Semantic Segmentation (AAAI 2024 · similarity 0.88)
- X-3D: Explicit 3D Structure Modeling for Point Cloud Recognition (CVPR 2024 · similarity 0.88)
- VSRD: Instance-Aware Volumetric Silhouette Rendering for Weakly Supervised 3D Object Detection (CVPR 2024 · similarity 0.87)
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.
