The Wolfram Language provides built-in functions for generating standard structure matrices and convolution kernels in any number of dimensions, in a form that can be used directly in image processing, linear algebra, or other applications.

Shape Matrices

DiskMatrix — BoxMatrix — CrossMatrix — DiamondMatrix —

Convolution Kernels

GaussianMatrix — Gaussian and Gaussian derivatives

ShenCastanMatrix — exponential and exponential derivatives

SavitzkyGolayMatrix — Savitzky–Golay smoothing and derivative kernel

GaborMatrix — Gabor kernel

IdentityMatrix  ▪  DiagonalMatrix

ConstantArray — constant array

CenterArray — embed an array at the center of another array

SparseArray — arbitrary sparse array

ArrayFlatten — form a matrix from tiles

ArrayReshape — reshape the array to new dimensions

ArrayPad — add padding to an array

ArrayResample — resample an array to new dimensions

ArrayFilter — apply a function to array neighborhood blocks