unravel.image_tools.smooth module#
Use img_smooth (smooth) from UNRAVEL to smooth an image.nii.gz and save it.
Note
Sigma is the Gaussian standard deviation.
By default, sigma is interpreted in voxels (scipy.ndimage.gaussian_filter).
With –mm, sigma is interpreted in millimeters (FSL fslmaths -s).
This matches conventions used elsewhere in UNRAVEL (e.g.,
reg: voxels;vstats: mm).Smoothing is performed in 3D.
- Inputs:
path/img.nii.gz or glob pattern(s)
- Outputs:
path/img_s<sigma><mm|vx>.nii.gz
Usage:#
img_smooth [-i path/img.nii.gz or glob pattern(s)] [-s sigma] [-m] [-v]
- unravel.image_tools.smooth.smooth_w_sigma_in_mm(img_path, sigma_mm, output_path)[source]#
Smooth an image with a Gaussian kernel using FSL’s fslmaths -s, which expects sigma in millimeters.
Parameters:#
- img_pathstr or Path
Path to the input image (e.g., .nii.gz).
- sigma_mmfloat
Standard deviation of the Gaussian kernel in millimeters.
- output_pathstr or Path
Path to save the smoothed image (e.g., .nii.gz).
- unravel.image_tools.smooth.smooth_w_sigma_in_voxels(img_path, sigma_vx, output_path)[source]#
Smooth an image with a Gaussian kernel using scipy.ndimage.gaussian_filter, which expects sigma in voxels.
Parameters:#
- img_pathstr or Path
Path to the input image (e.g., .nii.gz).
- sigma_vxfloat
Standard deviation of the Gaussian kernel in voxels.
- output_pathstr or Path
Path to save the smoothed image (e.g., .nii.gz).
- unravel.image_tools.smooth.smooth_image(img_path, sigma, mm=False)[source]#
Smooth an image with a Gaussian kernel, interpreting sigma in either millimeters or voxels.
Parameters:#
- img_pathstr or Path
Path to the input image (e.g., .nii.gz).
- sigmafloat
Standard deviation of the Gaussian kernel.
- mmbool, optional
If True, interpret sigma in millimeters and use FSL’s fslmaths -s. If False, interpret sigma in voxels and use scipy.ndimage.gaussian_filter. Default is False.