Medical Imaging Operators

Authors: Holoscan SDK Team (NVIDIA)
Supported platforms: x86_64, aarch64
Language: Python
Last modified: August 5, 2025
Latest version: 1.1.0
Minimum Holoscan SDK version: 1.0.3
Tested Holoscan SDK versions: 2.2.0, 3.2.0
Contribution metric: Level 2 - Trusted

Medical image processing and inference operators.

Overview

This set of operators accelerate the development of medical imaging AI inference application with DICOM imaging network integration by providing the following,

• application classes to automate the inference with MONAI Bundle as well as normal TorchScript models
• classes to load supported AI model from files to detected devices, GPU or CPU
• classes to parse runtime options and well-known environment variables
• DICOM study parsing and selection classes, as well as DICOM instance to volume image conversion
• DICOM instance writers to encapsulate AI inference results in these DICOM OID,
• DICOM Segmentation
• DICOM Basic Text Structured Report
• DICOM Encapsulated PDF
• Surface mesh generation and storage in STL format
• Visualization with Clara-Viz integration, as needed

Requirements

This set of operators depends on Holoscan SDK Python package, as well as directly on the following,

• highdicom
• monai
• nibabel
• numpy
• numpy-stl
• Pillow
• pydicom
• PyPDF2
• scikit-image
• SimpleITK
• torch
• trimesh
• typeguard

Notices

Many of this set of operators are Derivative Works of MONAI Deploy App SDK under its Apache-2.0 license, and Nvidia employees have been the main contributors to MONAI Deploy App SDK.

The dependency packages' licences can be viewed at their respective links as shown in the above section.