Texture Filters

This group contains algorithms that reveal texture attributes.

MeanFilter2d: Computes, for each pixel of a two-dimensional image, the mean value of its neighborhood.
MeanFilter3d: Computes, for each voxel of a three-dimensional image, the mean value of its neighborhood.
VarianceFilter2d: Computes, for each pixel of a two-dimensional image, the variance value of its neighborhood.
VarianceFilter3d: Computes, for each voxel of a three-dimensional image, the variance value of its neighborhood.
SkewnessFilter2d: Computes, for each pixel of a two-dimensional image, the skewness value of its neighborhood.
SkewnessFilter3d: Computes, for each voxel of a three-dimensional image, the skewness value of its neighborhood.
KurtosisFilter2d: Computes, for each pixel of a two-dimensional image, the kurtosis value of its neighborhood.
KurtosisFilter3d: Computes, for each voxel of a three-dimensional image, the kurtosis value of its neighborhood.
ContrastFilter2d: Computes, for each pixel of a two-dimensional image, the contrast value of its neighborhood.
ContrastFilter3d: Computes, for each voxel of a three-dimensional image, the contrast value of its neighborhood.
VariationFilter2d: Computes, for each pixel of a two-dimensional image, the variation value of its neighborhood.
VariationFilter3d: Computes, for each voxel of a three-dimensional image, the variation value of its neighborhood.
EnergyFilter2d: Computes, for each pixel of a two-dimensional image, the energy value of its neighborhood.
EnergyFilter3d: Computes, for each voxel of a three-dimensional image, the energy value of its neighborhood.
EntropyFilter2d: Computes, for each pixel of a two-dimensional image, the entropy value of its neighborhood.
EntropyFilter3d: Computes, for each voxel of a three-dimensional image, the entropy value of its neighborhood.
AutocorrelationFilter: Computes the cross-correlation of an image by itself.
LocalAutocorrelationFilter2d: Computes the cross correlation of a two-dimensional image by itself in a given window size.
ImageCurvature2d: Computes local gray level curvature in a two-dimensional image.
ImageCurvature3d: Computes local gray level curvature in a two-dimensional image.
StructureEnhancementFilter2d: Enhances structures of interest in a two-dimensional image using a multi-scale analysis.
StructureEnhancementFilter3d: Enhances structures of interest in a three-dimensional image using a multi-scale analysis.
ScaleSpaceLogFilter2d: This filter applies a Laplacian of Gaussian (LoG) filter at different scales, to enhance circular or elliptical (blob-like) objects.

The Texture Filters group contains filters that can be used to reveal the features of a texture; for instance, by replacing a given pixel intensity by a first order statistic of its neighborhood.

First Order Statistics Filters

Most of the algorithms exposed here deal with first order statistics. They create a result image where a pixel value is only a function of the corresponding pixel in the initial image and its neighbourhood with a user-defined size. To calculate these values, local histograms are computed (arrays of pixel counts per value in the neighbourhood). In the equations given for each algorithm, $p(n)$ corresponds to the probability to have a pixel with value $n$. It is simply the number of those pixels divided by the total number of pixels in the neighbourhood.