How to Run & View AI Model Outputs: Using TotalSegmentator

Overview

This guide shows you how to run the TotalSegmentator AI model in the 3DICOM Viewer and view its segmentation outputs. You’ll also find key recommendations for achieving optimal AI segmentation results.

What is TotalSegmentator?

TotalSegmentator automatically identifies and labels anatomical structures in compatible CT and MRI studies.

These segmented structures can be explored using Annotiva for interactive anatomical descriptions and learning support.

Before you begin

Ensure your imaging is high-quality and covers the complete anatomy.
Series with missing slices or poor image quality may lead to incomplete or inaccurate segmentations.
To quickly check, preview the series in 3D in 3DICOM: if the anatomy is clear and complete, the AI segmentation results will likely be reliable.

Run the TotalSegmentator AI model

Open the AI Model Marketplace

Select the brain-shaped icon from the toolbar to open the AI Model Marketplace.

The Marketplace lists all AI models available to your account, along with your current credit balance.

Important: TotalSegmentator is automatically selected, as it’s currently the only available model. Although some AI models may require credits in the future, TotalSegmentator is currently free to run.

Select your study and series

Choose the study you wish to process.
If the study contains multiple imaging series, select the ones you want to include. All series are selected by default—deselect any you don’t want to process. Series names and slice counts correspond to the previews in the Study Browser on the left side of the interface.
Click Run AI Model to start processing. The model will run in the background. Depending on the number and size of selected series, this may take up to twenty (20) minutes.

Note: While processing, you can continue working within the viewer or elsewhere in the platform.

Monitor processing and notifications

When processing completes, you’ll receive a notification in the upper right corner of the interface.
Click the notification to open the processed study in the Viewer Segmentation workspace. (This workspace is also accessible any time via the workspace icon in the Viewer toolbar.)

View and interact with segmentation outputs

AI-generated segmentation outputs (SEG files) will appear in the Study Browser (located on the left of the interface) under the original imaging series. To load the segmentations for a series, double-click the relevant SEG file.
Once loaded, the AI-generated anatomical segments display in the Segmentation panel on the right. Browse and interact with the identified structures as needed. Segmentation results are available in both 2D and 3D viewports, with Annotiva support (denoted by the magic wand icon in the toolbar).

Tip: You can revisit your segmentation results at any time by reopening the study from Scans & Records and loading the SEG file from the Study Browser.

Result

You can now successfully run the TotalSegmentator AI model in the 3DICOM Viewer, select the appropriate study and series, monitor processing, and load the completed segmentation results for review and further analysis.