wmh-segmentation

What problem does it solve?

This skill provides a safe, reproducible interface to perform automated white matter hyperintensity (WMH) segmentation on structural MRI data by wrapping the MARS-WMH nnU-Net Docker container. It handles input validation, dependency checks, GPU readiness, workspace preparation, and execution orchestration to deliver a ready-to-use segmentation mask with clear reporting.

Core Features & Use Cases

• Input validation for FLAIR and T1w NIfTI images (no contrast) and co-registration checks.
• Docker-based WMH segmentation using the official MARS-WMH nnU-Net container with GPU support.
• End-to-end orchestration: dependency checks, workspace setup, execution, and result reporting for reproducible runs.
• Use case: a researcher wants to segment WMH on a FLAIR+T1w pair and obtain a NIfTI mask with a deterministic workflow.

Quick Start

Provide paths to your FLAIR and T1w NIfTI images when prompted to run the WMH segmentation workflow.