Repository: https://github.com/isaac-for-healthcare/i4h-tutorials

Synthetic data generation pipelines to bridge healthcare robotic data gap, from anatomies to hospitals to robots

Isaac for Healthcare Synthetic Data Generation

This repository contains tutorials and reusable components for customizing and extending Isaac for Healthcare workflows with your own assets, data, and environments. The content is organized around digital twin pipelines (patient, hospital, robot) and surgical robotic simulation and video generation.

Available Pipelines

Patient Digital Twin Turn clinical data into simulation-ready 3D assets — synthetic medical images, CT/MR to USD conversion, and style augmentation

Hospital Digital Twin Build simulation-ready environments — environment creation, robot rigging, teleoperation, recording, and visual style augmentation

Robot Digital Twin Bring your own robot into the pipeline — custom URDF/CAD, rigging in Isaac Sim, and integration with the simulation

Surgical Robotic Video Generator Bridge Cosmos world models with downstream policy and use Cosmos-Transfer for data augmentation and policy generalizability

Surgical Robotic Generative Physics Simulator Post-train Cosmos on custom surgical robotics datasets for policy evaluation and synthetic data generation