A brain segmentation tool
Project description
Brain Seg
A brain segmentation tool using machine learning
Under development
This software has NOT been approved for medical analysis/diagnosis. Please use other approved software for medical use.
Models
SavedModels for use can be found at Google Drive.
Models were trained using the Mindboggle-101 data:
Klein, Arno, 2016, "Mindboggle-101 manually labeled individual brains", https://doi.org/10.7910/DVN/HMQKCK, Harvard Dataverse, V2
A Klein, SS Ghosh, FS Bao, J Giard, Y Hame, E Stavsky, N Lee, B Rossa, M Reuter, EC Neto, A Keshavan. 2017. Mindboggling morphometry of human brains. PLoS Computational Biology 13(3): e1005350. doi:10.1371/journal.pcbi.1005350 (bioRxiv 2016 preprint)
Options
positional arguments:
filepath Path of file to segment
optional arguments:
-h, --help show this help message and exit
-s SEGMENT, --segment Segments using provided SavedModel (Directory path required)
--segment-precision Float32 or float64 {32,64}
--output OUTPUT, -o Output filepath
--output-channels Number of channel in output image {1,3}
Usage Examples
# Segments a sample nifti scan
$ brainseg sample.nii.gz -s sampleSavedModel -o output.nii.gz
Model Training Graphs
SavedModel_210303
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.