How to apply multiple convolutions in one step?

I have a matrix of signals where each row corresponds to a signal and I also have a corresponding matrix of filters where each row is a filter.

Thus, each signal has a filter associated with it. How do I apply the corresponding convolutions at once?

Answer 1

[MathWorks Support Team]

One way to solve this problem is to convert the filters and signals into cell arrays and make use of 'cellfun' operation to apply row by row convolution.

Assume:

A is the signal matrix where the rows are different signals.

B is the filter matrix with 2 filters

A = [[1 0 1];[1 0 1]];
B = [[2,7];[2,7]];

Convert these matrices into a cell array for faster computation

cellA = num2cell(A, 2)
cellB = num2cell(B,2)

Perform convolution on each row using 'cellfun':

result = cellfun(@(A,B) conv(A, B), cellA,cellb, 'UniformOutput', false);

The last function applies convolution to each row of A with its corresponding filter from B.

'cellfun' applies the desired function (convolution here) to each element of the cell array. More details about 'cellfun' can be found here:

https://www.mathworks.com/help/matlab/ref/cellfun.html