This function is not yet fully documented. This is a transcript of the text-formatted help.
bbconvn Create a convoluting blackbox operator. [BB,DIM2]=bbconvn(DIM1,K) creates a black-box operator that convolves a hypercube of dimensions DIM1 with the kernel K. The hypercube is represented as a column vector. The kernel K must have as many dimensions as the length of DIM1. The resulting hypercube has the dimensions given by DIM2. Several kernels of high quality is available using bbkernel. [BB,DIM2]=bbconvn(DIM1,K,S) controls the output size: 'full' - (default) returns the full N-D convolution: DIM2=DIM1+size(K)-1. 'same' - returns the central part of the convolution: DIM2=DIM1. 'valid' - returns only the part of the result that can be computed without assuming the input is zero-padded: DIM2=max(DIM1-size(K)+1,0) Example: Create an operator which applies a Gaussian filter to a 100-by-100 image, producing an image of the same size. K=bbkernel([10,10],'gauss',.2); BB=bbconvn([100,100],K,'same'); See also bbkernel, .