Interactive Segmentation based on the Existing Baseline

About

Editor:

mateo-thorner

Contact email:

mateo.thorner@sia-partners.com

Image Version:

c81b31be-5d40-4111-ab21-d1f6ed14ee0f — Aug. 19, 2026

Model Version:

5921280e-c540-46a2-ab30-5a826386f28b — Aug. 19, 2026

Summary

The algorithm basically is the same as the Baseline Proposed for task 2.

The only modification that we made was to treat only .nii.gz instead of .mha to save disk space during the preprocessing phases.

Mechanism

The algorithm is the same as the baseline already existing for the task2

It consists of 2 mains phases : - The first phase does a global segmentation of the bones -> Anatomic part - The second phase does multiple passes on every fragments of the anatomy pointed by the clicks and segments it properly. -> Fragments part

Interfaces

This algorithm implements all of the following input-output combinations:

Inputs Outputs

	Inputs	Outputs
1	Peripelvic Fracture CT Slug `peripelvic-fracture-ct` Description CT scan of a fracture around the pelvis. Kind Image Read from `/input/images/peripelvic-fracture-ct/<uuid>.mha` or `/input/images/peripelvic-fracture-ct/<uuid>.tif` Peripelvic Fracture CT Peripelvic Fragment Clicks Slug `peripelvic-fragment-clicks` Description Points of interest that represent clicks on fragments around the pelvis. Each point is categorized by one of four types specified in the "type" field: "Center of Mass Points of Interest", "Uniformly Sampled Points of Interest", "Boundary Internal Margin Points of Interest", or "Euclidean Distance Transform Points of Interest". Points are grouped by fragment, with each group identified by its "fragment_id" field. Kind Multiple points Read from `/input/peripelvic-fragment-clicks.json` Example file Download an example. Example value { "name": "Center of Mass Points of Interest", "type": "Multiple points", "points": [ { "name": "Center of Mass Point 1", "point": [ 96.0, 79.8, 0.5 ] }, { "name": "Center of Mass Point 2", "point": [ 130.1, 115.5, 0.5 ] } ], "version": { "major": 1, "minor": 0 } } Peripelvic Fragment Clicks	Peripelvic Fracture CT Segmentation Slug `peripelvic-fracture-ct-segmentation` Description Segmentation of a pelvic fracture (0=background, 1-50 sacrum fragments, 50-100 left iliac fragments, 101-150 right iliac fragments, 151-200 femur fragments). Kind Segmentation Write to `/output/images/peripelvic-fracture-ct-segmentation/<uuid>.mha` or `/output/images/peripelvic-fracture-ct-segmentation/<uuid>.tif` Peripelvic Fracture CT Segmentation

Validation and Performance

Left empty by the Algorithm Editors

Challenge Performance

Date	Challenge	Phase	Rank
Aug. 19, 2026	PENGWIN2026	Task 2: PENGWIN-Interact-Final Test	31

Uses and Directions

This algorithm was developed for research purposes only.

Warnings

Left empty by the Algorithm Editors

Common Error Messages

Left empty by the Algorithm Editors

Information on this algorithm has been provided by the Algorithm Editors, following the Model Facts labels guidelines from Sendak, M.P., Gao, M., Brajer, N. et al. Presenting machine learning model information to clinical end users with model facts labels. npj Digit. Med. 3, 41 (2020). 10.1038/s41746-020-0253-3