FastSurferCNN Processing & Visualization with InVesalius

This repository provides a streamlined workflow for running FastSurferCNN for brain MRI segmentation and visualizing the results in InVesalius.

📌 Overview

• FastSurferCNN: A deep-learning-based tool for fast whole-brain segmentation of MRI scans.
• InVesalius: A free, open-source software for 3D reconstruction of medical imaging data (DICOM/NIfTI).

This guide walks through:
👉 Running FastSurferCNN on an MRI scan
👉 Converting the output into a visualizable format
👉 Loading and analyzing the results in InVesalius

🚀 Installation

1️⃣ Clone this repository

git clone https://github.com/your-username/your-repo.git
cd your-repo

2️⃣ Install FastSurferCNN

git clone https://github.com/Deep-MI/FastSurfer.git
cd FastSurfer
pip install -r requirements.txt  # Install dependencies

3️⃣ Install InVesalius

Download and install InVesalius from:
👉 https://invesalius.github.io/download/

🏃 Running FastSurferCNN on MRI Data

Make sure your input MRI scan is in NIfTI format (.nii or .nii.gz).

python FastSurferCNN.py --i input.nii.gz --o output_dir

Output Files (in output_dir/):

• aseg.auto.mgz → Segmentation mask
• mri_orig.mgz → Processed input image

💡 Convert .mgz to .nii.gz for InVesalius compatibility:

mri_convert output_dir/aseg.auto.mgz output_dir/aseg.nii.gz

🔦 Visualizing in InVesalius

1️⃣ Open InVesalius
2️⃣ Click "File" → "Import Medical Images"
3️⃣ Select aseg.nii.gz
4️⃣ Adjust opacity & color settings for better visualization

🟢 Optional: Overlay segmentation on original MRI

• Load mri_orig.nii.gz
• Set segmentation mask transparency

🛠 Troubleshooting

🛠 FastSurferCNN not detecting GPU?
Try running on CPU:

python FastSurferCNN.py --i input.nii.gz --o output_dir --no_cuda

🛠 InVesalius not displaying segmentation?

• Ensure you converted .mgz to .nii.gz
• Adjust contrast & opacity settings

📜 References

• FastSurfer GitHub
• InVesalius Official Website

✨ Contributions & Issues

Feel free to fork, submit PRs, or report issues. 🚀