FMRIB's Automated Segmentation Tool (FAST)¶
Library: FSL | Docker Image: brainlife/fsl
Function¶
Segments brain images into gray matter, white matter, and CSF using a hidden Markov random field model with integrated bias field correction.
Modality: Brain-extracted T1-weighted 3D NIfTI volume.
Typical Use: Tissue probability maps for normalization, VBM studies, or masking.
Key Parameters¶
-n (number of tissue classes, default 3), -t (image type: 1=T1, 2=T2, 3=PD), -B (output bias field), -o (output basename)
Key Points¶
Input must be brain-extracted. Outputs partial volume maps (*_pve_0/1/2) for each tissue class. Use -B to get estimated bias field.
Inputs¶
| Name | Type | Required | Label | Flag |
|---|---|---|---|---|
input |
File |
Yes | ||
output |
string |
Yes | Output filename prefix | -o |
nclass |
int |
No | Number of tissue classes | -n |
iterations |
int |
No | Number of iterations during bias-field removal | -I |
lowpass |
double |
No | bias field smoothing extent (FWHM) in mm | -l |
image_type |
int |
No | Image type (e.g. 1="T1", 2="T2", 3="PD") | -t |
fhard |
double |
No | initial segmentation spatial smoothness (during bias field estimation) | -f |
segments |
boolean |
No | Outputs a separate binary segmentation file for each tissue type | -g |
bias_field |
boolean |
No | Outputs estimated bias field | -b |
bias_corrected_image |
boolean |
No | Outputs bias-corrected image | -B |
nobias |
boolean |
No | Do not remove bias field | -N |
channels |
int |
No | Number of channels to use | -S |
initialization_iterations |
int |
No | initial number of segmentation-initialisation iterations | -W |
mixel |
double |
No | spatial smoothness for mixeltype | -R |
fixed |
int |
No | number of main-loop iterations after bias-field removal | -O |
hyper |
double |
No | segmentation spatial smoothness | -H |
manualseg |
File |
No | Manual segmentation file | -s |
probability_maps |
boolean |
No | outputs individual probability maps | -p |
priors |
record(priors) |
No |
Accepted Input Extensions¶
- input:
.nii,.nii.gz - manualseg:
.nii,.nii.gz - initialize_priors:
.mat
Outputs¶
| Name | Type | Glob Pattern |
|---|---|---|
segmented_files |
File[] |
$(inputs.output)_seg.nii.gz, $(inputs.output)_pve_*.nii.gz, $(inputs.output)_mixeltype.nii.gz, $(inputs.output)_pveseg.nii.gz |
output_bias_field |
File |
$(inputs.output)_bias.nii.gz |
output_bias_corrected_image |
File |
$(inputs.output)_restore.nii.gz |
output_probability_maps |
File[] |
$(inputs.output)_prob_*.nii.gz |
output_segments |
File[] |
$(inputs.output)_seg_*.nii.gz |
log |
File |
$(inputs.output).log |
Output Extensions¶
- segmented_files:
.nii.gz - output_bias_field:
.nii.gz - output_bias_corrected_image:
.nii.gz - output_probability_maps:
.nii.gz - output_segments:
.nii.gz
Enum Options¶
image_type: 1, 2, 3
Docker Tags¶
Available versions: latest, 6.0.4-patched2, 6.0.4-patched, 6.0.4, 6.0.4-xenial, 5.0.11, 6.0.0, 6.0.1, 5.0.9
Categories¶
- Structural MRI > FSL > Tissue Segmentation