Brain Hippocampus Segmentation

This program can quickly segment (<1min) the Hippocampus of raw brain T1 images.

It relies on a Convolutional Neural Network pre-trained on thousands of images from multiple large cohorts, and is therefore quite robust to subject- and MR-contrast variation. For more details on its creation, refer the corresponding manuscript at http://dx.doi.org/10.1016/j.media.2017.11.004

This version is a PyTorch port of the original Theano model. This version has less dependencies and is notably faster thanks to pytorch multithreading. While the hippocampal segmentation model is exactly the same as described in the paper, the pre- and post-processing steps had been improved, and thus, results may slightly differ. The original repo is still available at https://github.com/bthyreau/hippodeep

This program requires Python 3, with the PyTorch library, version > 1.0.0.

No GPU is required

Tested on Linux CentOS 6.x/7.x, Ubuntu 18.04 and MacOS X 10.13, using PyTorch versions 1.0.0 to 1.4.0. In addition, Windows compatibility patches by Bernd Foerster are available at https://github.com/bfoe/hippodeep_pytorch

Just clone or download this repository.

In addition to PyTorch, the code requires scipy and nibabel.

The simplest way to install from scratch is maybe to use a Anaconda environment, then

• install scipy (conda install scipy or pip install scipy) and nibabel (pip install nibabel)
• get pytorch for python from https://pytorch.org/get-started/locally/. CUDA is not necessary.

To use the program, simply call:

deepseg1.sh example_brain_t1.nii.gz.

To process multiple subjects, pass them as multiple arguments. deepseg1.sh subject_*.nii.gz.

The resulting segmentations should be stored as example_brain_t1_mask_L.nii.gz (or R for right) and example_brain_t1_brain_mask.nii.gz. The mask volumes (in mm^3) are stored in a csv file named example_brain_t1_hippoLR_volumes.csv. If more than one input was specified, a summary table named all_subjects_hippo_report.csv is created.