A pot-pourri of utilities related to the processing and analysis of segmentation, parcellation files, reorientation, translation between common parcellation atlases
Allows either to compare two segmentation or pairs of segmentation and reference as specified per folder.
Output summary as a csv file listing for each pair of reference / segmentation image the measures of interest
List of measures includes:
- Volume of each,
- Number of true positives, true negatives, false positives, false negatives
- Outline error divided into false positives and negatives
- Detection error divided in false positives and negatives
- Dice score
- Jaccard score
- Hausdorff distance
- Average distance
- Volume difference
If specified, with the -save_map option, the maps of different types of error is saved as well
Allows the computation of mask based statistics over an image from distribution features to texture features
Returns a csv file with all the required measures for all the files listed with pairs of mask and image
Measures include:
- Distribution features
- Mean, min, max, std
- 1st, 5th, 25th, 50th, 75th, 95th, 99th percentile
- Skewness, Kurtosis
- Shape features
- Volume, surface, surface to area volume, compactness
- Border (external, internal)
- Texture features
- Haralick features
Enable the translation between different parcellation labelling.
Notably available GIF (Cardoso et al 2015) to Freesurfer labelling
With the option mask_mapping enables the relabeling over a mask
With the subcommand split_label, enables the split of a given label according to a direction based on ellipsoid decomposition
Different subcommand application:
- parsing: Allows the parsing of xml files (specifically GIF output) into a database with all labels and hierarchically organised aggregates
- seg_aggregate: Allows for the aggregation of multiple region to create a coarser segmentation from a parcellation files. Useful to get whole lobes or specific sublobes.
- database_fromparc: When no xml file is available but a list of nifti parcellation files, create the database using the hierarchy file.
Gather small utility function to reorient nifti files, transform from flirt to niftyreg affine matrix (similarly to nifty reg) and from niftyreg to flirt affine matrix.
Different subcommands are
- checks: Check anisotropy, and acquisition of provided images.
- flirtnr: Performs the transformation of an affine matrix from FLIRT into NiftyReg format. Requires the floating image, reference image and the matrix file to modify
- nrflirt: Performs the transformation of an affine matrix from niftyreg into flirt affine matrix. Requires the floating image, reference image and the matrix file to modify
- reor: Reorient the supplied nifti image into the desired orientation