domain-3d
Skill8 starsChanged 4 months agoMITView on GitHub
Use for 3D and point cloud tasks: point cloud classification, 3D semantic or instance segmentation, 3D object detection from LiDAR, and depth or mesh analysis, including autonomous-driving and indoor-scan settings. Picks the right 3D backbone, representation, and scene-level split and metric. Triggers on 'point cloud', '3D', 'LiDAR', 'mesh', 'depth', '3D detection', '3D segmentation', 'ScanNet', 'KITTI', 'ModelNet', 'autonomous driving perception'.
In this repository22 other files it gives its agents.
- accuracy-improvement-loopSkill
- data-rigor-and-leakageSkill
- dataset-profilingSkill
- deployment-explainabilitySkill
- domain-anomaly-detectionSkill
- domain-audio-speechSkill
- domain-computer-visionSkill
- domain-generativeSkill
- domain-graphSkill
- domain-medical-imagingSkill
- domain-multimodalSkill
- domain-nlp-llmSkill
- domain-recommenderSkill
- domain-reinforcement-learningSkill
- domain-tabularSkill
- domain-time-seriesSkill
- domain-videoSkill
- literature-reviewSkill
- ml-research-methodologySkill
- notebook-deliverySkill
- rigorous-evaluationSkill
- training-optimizationSkill
The file
Files
3D and Point Cloud - Method Selection
Choose a representation first: raw points, or sparse voxels for large scenes. Normalize and center the cloud.
Decision table
| Task | Recommended (recent, 2022-2025) | Notes |
| Point cloud classification | PointNeXt or Point Transformer v3; PointNet++ as a baseline | ModelNet-style. Metric is overall and mean-class accuracy. |
| Semantic segmentation (indoor or outdoor) | Point Transformer v3 (SOTA 2024) or sparse convolution (MinkowskiNet) | ScanNet, S3DIS. Metric is mIoU. |
| 3D object detection (driving) | sparse-voxel detectors (CenterPoint, VoxelNeXt) or transformer detectors | KITTI, nuScenes. Metric is mAP and NDS. |
| Efficiency-critical or very large clouds | Point Mamba (linear-time, 2024-2025) or sparse convolution | trade a little accuracy for speed and memory. |
Cross-cutting practice
- Augment with rotation, jitter, scaling, and random point dropout. Voxelize for large outdoor scenes.
- Leakage: split by scene or scan, never by cropped block within the same scene.
- Metrics: overall accuracy for classification, mIoU for segmentation, mAP and NDS for detection.
- Explainability: per-point saliency; highlight the points that drove the prediction.
- Improve accuracy: use
accuracy-improvement-loop; evaluate withrigorous-evaluation.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.