itk_diffusion

apply anisotropic diffusion filtering to a volume

itk_diffusion <input> <output> [options]

DESCRIPTION

itk_diffusion applies anisotropic diffusion filtering to an image volume using the Insight Toolkit (ITK). This edge-preserving smoothing technique reduces noise while retaining important structural boundaries. It is particularly useful as a preprocessing step for segmentation and registration of medical images.

Two diffusion models are available. The default gradient-magnitude model (Perona-Malik) uses the gradient magnitude to control the diffusion rate, reducing smoothing near strong edges. The alternative curvature model (--curvature) uses curvature-driven diffusion, which better preserves thin structures and sharp corners.

The strength of the filtering is controlled by the number of iterations (--iter) and the conductance parameter (--conduct). Higher conductance values allow more diffusion across edges; lower values restrict smoothing to homogeneous regions.

OPTIONS

--verbose

Print verbose information during processing.

--clobber

Overwrite the output file if it already exists.

--iter n

Number of diffusion iterations. More iterations produce stronger smoothing. Default is 5.

--conduct val

Conductance parameter controlling edge sensitivity. Higher values smooth more aggressively across edges. Default is 1.0.

--curvature

Use the curvature anisotropic diffusion model instead of the default gradient magnitude model. The curvature model better preserves thin structures and sharp corners.

EXAMPLES

Apply default gradient-based anisotropic diffusion:

itk_diffusion input.mnc smoothed.mnc

Apply 10 iterations with low conductance for strong edge preservation:

itk_diffusion input.mnc smoothed.mnc --iter 10 --conduct 0.5

Use the curvature diffusion model:

itk_diffusion input.mnc smoothed.mnc --curvature --iter 8 --conduct 1.5

Overwrite an existing output:

itk_diffusion noisy.mnc filtered.mnc --iter 20 --conduct 2.0 --clobber

AUTHOR

Vladimir S. Fonov - Montreal Neurological Institute, McGill University.

COPYRIGHTS

Copyright (C) Vladimir S. Fonov, McConnell Brain Imaging Centre, McGill University.

SEE ALSO

itk_laplace, itk_laplacian_sharpening, mincblur, itk_g_morph