Research map / Point clouds / Segmentation and detection
3DSSD: Point-Based 3D Single Stage Object Detector
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace D-FPS-only sampling with fusion sampling (ablation) | 3DSSD · KITTI, Car moderate | AP: 70.4 -> 79.4 | Helped |
| Add D-FPS to F-FPS sampling (ablation) | 3DSSD · KITTI, Car moderate | AP: 76.7 -> 79.4 | Helped |
| Use F-FPS instead of D-FPS (ablation) | 3DSSD · KITTI, Car moderate | AP: 70.4 -> 76.7 | Helped |
| Shift F-FPS representative points to generate candidate points (ablation) | 3DSSD with 3D center-ness assignment · KITTI, Car moderate | AP: 43.0 -> 79.4 | Helped |
| Use 3D center-ness rather than IoU assignment (ablation) | 3DSSD with shifting · KITTI, Car moderate | AP: 78.1 -> 79.4 | Helped |
| Use 3D center-ness rather than mask assignment (ablation) | 3DSSD with shifting · KITTI, Car moderate | AP: 77.3 -> 79.4 | Helped |
| Use feature-distance FPS with λ=1.0 rather than D-FPS (ablation) | 3DSSD · nuScenes, 1024 representative points | points recall: 65.9% -> 89.2% | Helped |
| Combine spatial and feature distances in F-FPS (λ=1.0) rather than feature distance only (λ=0.0) (ablation) | 3DSSD · nuScenes, 1024 representative points | points recall: 83.5% -> 89.2% | Helped |
| Use 3DSSD instead of PointPillars | 3D object detection · nuScenes | mAP: 29.5 -> 42.66 | Helped |
| Use 3DSSD instead of PointPillars | 3D object detection · nuScenes | NDS: 44.9 -> 56.4 | Helped |
| Use 3DSSD instead of STD | 3D object detection · KITTI test, Car moderate | AP3D (%): 79.71 -> 79.57 | Hurt |
| Use 3DSSD instead of STD | 3D object detection · KITTI | inference time (ms): 80 -> 38 | Helped |
Closest papers by meaning
- PVT-SSD: Single-Stage 3D Object Detector With Point-Voxel Transformer (CVPR 2023 · similarity 0.90)
- HVPR: Hybrid Voxel-Point Representation for Single-Stage 3D Object Detection (CVPR 2021 · similarity 0.89)
- SPGroup3D: Superpoint Grouping Network for Indoor 3D Object Detection (AAAI 2024 · similarity 0.87)
- Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detection (NeurIPS 2024 · similarity 0.87)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.87)
- GO-N3RDet: Geometry Optimized NeRF-enhanced 3D Object Detector (CVPR 2025 · 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.