Back to Adaptive Vision Library website

You are here: Start » Function Reference » Image » Image Local Transforms

Image Local Transforms

Select a function from the list below.

Icon Name Description / Applications Modules
CloseImage

Removes small dark structures from an image (or fills in bright ones) by applying consecutive dilation and erosion.


E.g. removal of the "pepper" component of salt-and-pepper noise.

FoundationLite
ConvolveImage

Computes a convolution of the input image with a user-specified mask.


Non-standard local transforms defined by the user.

FoundationLite
DifferenceOfGaussians

Applies difference of Gaussians on an image, i.e. computes difference of two Gaussian smoothed images.


Emphasizes high-frequency image features such as lines or patches / dots.

FoundationBasic
DilateAndErodeImage

Calculates dilation and erosion simultaneously for faster execution.

FoundationPro
DilateImage

Replaces each pixel with the maximum of pixels within a kernel.


Thickens bright features in an images and thins dark ones.

FoundationLite
DilateImage_AnyKernel

Replaces each pixel with the maximum of pixels within an arbitrary kernel.

FoundationPro
DilateImage_Mask

Replaces each pixel with the maximum of pixels within a small rectangular kernel.

FoundationLite
ErodeImage

Replaces each pixel with the minimum of pixels within a kernel.


Thins bright features in an image and thickens dark ones.

FoundationLite
ErodeImage_AnyKernel

Replaces each pixel with the minimum of pixels within an arbitrary kernel.

FoundationPro
ErodeImage_Mask

Replaces each pixel with the minimum of pixels within a small rectangular kernel.

FoundationLite
GradientDirAndPresenceImage

Computes an image of gradient directions mapped to the range from 1 to 255. Zero means "no edge".


For highly optimized analysis of gradient directions.

FoundationBasic
GradientImage

Computes a gradient image with smoothing operator of any size. The output pixels are signed.

FoundationLite
GradientImage_Mask

Computes a gradient image with a Sobel or Prewitt operator.

FoundationLite
GradientMagnitudeImage

Measures the strength of gradient at each pixel location.

FoundationLite
GradientMagnitudeImage_Signed

Computes an image of gradient for only selected direction.


For highly optimized analysis of gradient directions.

FoundationBasic
OpenImage

Removes small bright structures from an image (or fills in dark ones) by applying consecutive erosion and dilation.


E.g. removal of the "salt" component of salt-and-pepper noise.

FoundationLite
SmoothImage_Bilateral

Smooths an image while preserving sharp edges.

FoundationPro
SmoothImage_Deriche

Smooths an image using Deriche filter.


Approximation of the gaussian filter, which can be faster for large kernels.

FoundationLite
SmoothImage_DirAndPresence

Smooths the result of GradientDirAndPresenceImage.

FoundationPro
SmoothImage_Gauss

Smooths an image using a gaussian kernel.


Removal of gaussian noise from images.

FoundationLite
SmoothImage_Gauss_Mask

Smooths an image using a predefined gaussian kernel.


Removal of gaussian noise from images (fast).

FoundationLite
SmoothImage_Mean

Smooths an image by averaging pixels within a rectangular kernel.


Usually used for computing features related to local image "windows". Can be also used for noise removal, but Gauss is superior here.

FoundationLite
SmoothImage_Mean_AnyKernel

Smooths an image by averaging pixels within an arbitrary kernel.


Usually used for computing features related to local image "windows" having non-standard shape.

FoundationPro
SmoothImage_Mean_Mask

Smooths an image by averaging pixels within a small rectangular kernel.


This is a faster alternative to SmoothImage_Mean when the kernel is very small.

FoundationLite
SmoothImage_Median

Replaces each pixel with the median of pixels within a kernel.


Edge-preserving noise removal (but slow).

FoundationLite
SmoothImage_Median_Mask

Replaces each pixel with the median of pixels within a 3x3 rectangular kernel (faster).

FoundationLite
SmoothImage_Middle

Replaces each pixel with the average of maximum and minimum calculated within a kernel.


Useful for calculating per-pixel threshold values for image binarization.

FoundationLite
SmoothImage_Quantile

Replaces each pixel with a quantile of pixels within a kernel.


Edge-preserving noise removal (but slow).

FoundationLite
SmoothRegion_Mean

Smooths an region by averaging pixels within a rectangular kernel.


Usually used for computing features related to local image "windows". Can be also used for noise removal, but Gauss is superior here.

FoundationLite
StandardDeviationImage

Creates image of pixels' local standard deviations.

FoundationBasic