2-dimensional linear convolution by FFT

CONVOL2FFT computes a 2-dimensional linear convolution between an image and a filter.

CONVOL2FFT is a matlab function that returns the 2-dimensional linear convolution between a given image and a 2-dimensional impulse response of a filter. The filter must be a (2-dimensional) matrix. The image can be a matrix for a grayscale image and a 3D array for an RGB image. Maybe It is faster than IMFILTER, when size(h) > around 128. Typical syntaxes are as follows.
(1) x = convol2fft(x, h); where x * h is returned.
(2) [H,mns] = convol2fft(size(x),h); where H = DFT(h) and mns are returned.
(3) x = convol2fft(x,H,mns); where x * h is returned by using the given H = DFT(h) and mns.