Vss Deploy Detection Tracking 3d
Use when deploying or operating standalone RTVI-CV-3D / MV3DT multi-camera 3D tracking...
Install
npx skills add NVIDIA-AI-Blueprints/video-search-and-summarization --skill vss-deploy-detection-tracking-3d -g -a claude-code -y- Category
- Data and ML
- Updated
- 4d ago
- License
- Apache-2.0
- Context cost
- Mediumabout 3,000 tokens
- Runs scripts
- No, instructions only
What it does
Use when deploying or operating standalone RTVI-CV-3D / MV3DT multi-camera 3D tracking for calibrated MP4/file inputs and live RTSP streams: missing-calibration handoff to AMC skills, the 4-camera sample dataset, camera config, BEV Fusion, live OSD or saved grid/BEV outputs, bundled brokers, basic external MQTT/Kafka brokers, verification, and teardown. Trigger for generic MV3DT, RTVI-CV-3D, multi-view 3D tracking, multi-cam tracking, or sample MV3DT dataset requests. Explicit warehouse blueprint/profile MV3DT requests route to vss-build-vision-ai; single-camera 2D tracking routes to the 2D tracking or DeepStream skills. Not for full warehouse blueprint deployment, single-camera 2D tracking, camera calibration itself, or VSS summarization, Q&A, and RAG workflows.
VSS Deploy Detection And Tracking 3D
When to Use This Skill
Deploy the standalone RT-CV-3D MV3DT stack from services/rtvi/rt-cv-3d/rt-cv-mv3dt.
This is the default path for MV3DT / RTVI-CV-3D / multi-camera tracking requests.
Do not derive MV3DT services from the warehouse blueprint for this skill. Use
vss-build-vision-ai only when the user explicitly asks for warehouse MV3DT,
the warehouse blueprint, a bp_wh* profile, warehouse compose files, or the
combined warehouse application stack. When routing an explicit warehouse MV3DT
request, also state the boundary: generic MV3DT uses standalone RT-CV-3D, while
warehouse MV3DT uses warehouse/profile deployment. For single-camera 2D
detection or tracking, use the 2D tracking or DeepStream skills instead.
Public docs: https://docs.nvidia.com/vss/latest/object-detection-tracking.html.
Examples
Example operation prompts:
- "Deploy MV3DT on my calibrated four-camera MP4 dataset and save output."
- "Deploy MV3DT on the sample dataset."
- "Enable multi-camera tracking on the 4-cam example dataset."
- "Run RTVI-CV-3D on these RTSP streams after calibration."
- "Deploy multi-cam tracking; if there is no display, save the videos."
- "Use an external MQTT broker and external Kafka for this RT-CV-3D deployment."
- "Verify the standalone RT-CV-3D deployment and show output paths."
- "Tear down everything for standalone MV3DT."
Output Permissions
Keep output permissions scoped to the standalone runtime paths. If output writes fail, report the directory owner/mode, container user, and relevant logs instead of loosening permissions broadly.
What This Deploys
The standalone compose file is services/rtvi/rt-cv-3d/rt-cv-mv3dt/docker/compose.yml.
It deploys:
