How to compare matrix elements with their neighbors?

Hi, I have a matrix A (50x1). I have to compare each element of matrix A with its two neighbors. The element of the matrices are organized in a circular way and the neighbors definition is as follows: Element "i" has two neighbors: element "i+1" and element "i-1", For instance element 2 has neighbors 1 and 3, element 50 has neighbors 49 and 1, element 1 has neighbors 50 and 2. If the value of element "i" from matrix A is smaller than value of element "i+1" and element "i-1" from matrix A,then I build another matrix B which element "i" is equal to element "i" from A matrix.

Any advice how to write this is very appreciated.

Answer 1

[Image Analyst]

You can use imregionalmin() from the Image Processing Toolbox, but you'll have to pad your array

% Create sample data.
A = uint8(randi(255, 1, 50))
% Pad A and create temporary vector paddedA
% so that we don't alter A (we might need the original A later).
paddedA = [A(end), A, A(1)];
minLocations = imregionalmin(paddedA)
% Create B
B = zeros(1, length(paddedA));
% Assign in the values of A
B(minLocations) = paddedA(minLocations);
% Crop back down.
B = B(2:end-1)

[EB]

Thank you. Your code is doing exactly what I need.

Answer 2

[Guillaume]

That's extremely easy using simple indexing or, if you want to be a bit more fancy, using circshift

A = randi(20, 50, 1); %demo matrix

%using simple indexing
tocopy = A < [A(2:end); A(1)] & A < [A(end); A(1:end-1)];
%using circshift
tocopy = A < circshift(A, 1) & A < circshift(A, -1);

%now copy the selected values
B = zeros(size(A));  %create B first
B(tocopy) = A(tocopy);  %copy