How to find the dynamic central points of an image in matlab?

Question

I have a matrix named original and the respective plot is shown in the image below.

I would like to however get the center portion of the image in the form of respective rows into another variable as highlighted approximately in the next figure.

I looked up to find the center and got to know of the method of using centroid in regionprops. For example like the code sample below.

s = regionprops(original,'centroid');
centroids = cat(1, s.Centroid);
plot(centroids(:,1),centroids(:,2), 'b*')

This method of using centroid did not help me to obtain the entire central rows. Is there any alternative way of doing this?

Answer 1

One way to compute the average non-zero index in each column would be to directly compute it.

Edit: It's not clear what is meant by noise, but if I interpret this as you want to only consider the widest region per column then we can find the average of this region as follows.

avg = zeros(1,size(original,2));
for c = 1:size(original,2)
    % find largest contiguous region in the column
    sig = original(:,c);
    der = diff([false; sig; false]);
    upedge = find(der == 1);
    downedge = find(der == -1);
    [~,region_idx] = max(downedge-upedge);
    first = upedge(region_idx);
    last = downedge(region_idx)-1;

    % get the average index in the region
    avg(c) = (last + first) / 2;
end

% plot results
hold off;
imagesc(original); hold on;
plot(1:size(original,2), avg, '-r', 'LineWidth', 2);
set(gca, 'YDir', 'normal');
axis([1, size(original,2), 1, 250]);

Which, results in the following figure