3D Body Pose NIM — Container Card

Description

The 3D Body Pose NIM uses GPU-accelerated inference to estimate 3D human body poses and skeletal structures from video input. It processes video frames using an NVIDIA Triton Inference Server backend and returns pose-estimation results through a streaming gRPC API.

The container components are ready for commercial or non-commercial use.

License/Terms of Use

Use of this trial service is governed by the NVIDIA API Trial Terms of Service. Use of the GENMO model is governed by the NVIDIA Open Model License Agreement. Use of the Vitpose model is governed by the NVIDIA Open Model License Agreement, the Meta Dinov3 License and the SAM License; Built with Dinov3. Use of the SAM3DB-Body model is governed by the NVIDIA Open Model License Agreement and the SAM License.

Intended Use

• Real-time and offline 3D body pose estimation from video
• Video production, motion capture, sports analytics
• Accessibility applications
• Research and development in computer vision

Known Limitations

• Requires NVIDIA GPU, compute capability 8.0+ (Ampere and newer: sm80, sm86, sm89, sm90, sm100, sm120). T4/sm75 not supported.
• Multi-person pose estimation supported natively (tracked bounding boxes in, per-body 3D pose out); throughput scales with bodies/frame.
• Performance depends on input video resolution, bodies/frame, and GPU model
• Production 3D body pose model (AR SDK 3DBodyPoseEstimation, release 1.2.0) is integrated; not a dummy backend

Deployment Geography

Global

Release Date

Build.NVIDIA.com 09/30/2026 via https://build.nvidia.com/nvidia/body-pose NGC 09/30/2026 via https://catalog.ngc.nvidia.com/orgs/nim/teams/nvidia/containers/body-pose

Program Classes

The body_pose Container includes the AR SDK 3DBodyPoseEstimation (release 1.2.1), which ships as a bundle of the following models:

Deployment Details

Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA's hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

• Cloud: Hosted on build.nvidia.com (NV-GCP) via NVCF
• On-premise: Docker container with NVIDIA GPU runtime
• Ports: gRPC 8001, Triton HTTP 9000, Triton gRPC 9001
• Config: NIM_MODEL_PROFILE, NGC_API_KEY, NV_AI4M_MAX_CONCURRENCY_PER_GPU, NV_AI4M_MAX_INPUT_FILE_SIZE_BYTES, BODY_POSE_ENABLE_CONTACT, BODY_POSE_MAX_ACTIVE_REQUESTS — see nim-documentation getting-started/advanced-usage guides

Software Stack

Security Common Vulnerabilities and Exposures (CVEs)

Please review the Security Scanning tab on NGC to view the latest security scan results. For certain open-source vulnerabilities listed in the scan results, NVIDIA provides a response in the form of a Vulnerability Exploitability eXchange (VEX) document. The VEX information can be reviewed and downloaded from the Security Scanning tab.

Security

• gRPC transport unencrypted by default (NIM_SSL_MODE=disabled); TLS/mTLS opt-in

Ethical Considerations

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. Developers should work with their internal developer team to ensure these software components meet requirements for the relevant industry and use case and address unforeseen product misuse.

Please report quality, risk, security vulnerabilities, or NVIDIA AI concerns through the NVIDIA Vulnerability Disclosure Program.

Get Help

Getting started with the NIM

Deploying and integrating the NIM is straightforward thanks to our industry standard APIs. Visit the NIM container page on NGC for release documentation, deployment guides and more.

Enterprise Support

Get access to knowledge base articles and support cases or submit a ticket.