NIM for VISTA-3D

Type: Microservice Tags: NVIDIA, NIM, VISTA-3D, MONAI, medical imaging, segmentation, annotation, healthcare AI Related: NVIDIA-NIM, NVIDIA-MONAI-Toolkit, NVIDIA-Clara, NVIDIA-Clara-Viz, NIM-for-MAISI, NVIDIA-AI-Enterprise, NGC, TensorRT, Triton-Inference-Server, PyTorch Sources: https://docs.nvidia.com/nim/medical/vista3d/latest/overview.html, https://docs.nvidia.com/nim/medical/vista3d/latest/index.html, https://docs.nvidia.com/nim/medical/vista3d/latest/getting-started.html, https://docs.nvidia.com/nim/medical/vista3d/latest/advanced-usage.html, https://docs.nvidia.com/nim/medical/vista3d/latest/api-reference.html Last Updated: 2026-04-29

Summary

NVIDIA NIM for VISTA-3D packages VISTA-3D as a deployable NIM microservice for 3D medical imaging segmentation and annotation. Current NVIDIA docs describe VISTA-3D as an interactive foundation model for 3D medical imaging that supports segment-everything, class-based segmentation, and point-prompt workflows.

Detail

Purpose

VISTA-3D NIM gives researchers and medical imaging developers a containerized API for segmenting and annotating 3D medical images. It can accelerate anatomy exploration, organ/tumor mapping, and annotation workflows used to create ground-truth data.

Current scope

  • Segment-everything workflow for whole-body exploration.
  • Class-based segmentation for specific anatomy or disease analysis.
  • Point-prompt segmentation for interactive annotation.
  • API endpoints for health checks, model info, and /v1/vista3d/inference.
  • Input medical images are handled through NIfTI or NRRD-style workflows in current docs.
  • Multi-GPU load balancing through Triton Inference Server for high-volume workloads.
  • Documentation emphasizes research use, not clinical usage.

NVIDIA context

VISTA-3D NIM is the segmentation/annotation counterpart to NIM-for-MAISI in the medical imaging NIM graph. It connects NVIDIA-MONAI-Toolkit, NVIDIA-Clara, NVIDIA-Clara-Viz, and NIM deployment for medical imaging AI.

Connections

Source Excerpts

  • NVIDIA docs describe VISTA-3D as an interactive foundation model for 3D medical imaging.
  • The docs list segment-everything, class-based segmentation, and point-prompt segmentation as core workflows.

Resources