Separate Kernel in 1D kernels by Dirk-Jan Kroon

Decompose an arbitrary N dimensional filtering kernel into 1D kernels, for faster filtering (2d, gaussian, 3d)

[K1 KN ERR]=SeparateKernel(H)

FFT-based convolution by Bruno Luong

Discrete convolution using FFT method (conv, fft, convn)

convnfft(A, B, shape, dims, options)

convnfft_install

ndnanfilter.m by Carlos Adrian Vargas Aguilera

Filter/smoothing of multidimensional data with an specified window function, ignoring NaNs. (filter design, filter analysis, smooth)

ndnanfilter(X,HWIN,F,DIM,WINOPT,PADOPT,WNAN)