Find the median row of a binary column and replace the column with just the median row

I have some binary columns with ones and zeros. I would like to replace the places where there are multiple rows of ones with just one one in the median row index. I would like to do this for every column. For example:

1 0 1 1

1 0 1 1

1 1 1 0

0 1 1 0

0 1 0 0

becomes

0 0 0 1

1 0 1 0

0 0 0 0

0 1 0 0

0 0 0 0

The ones represent a curved y plot graph, that I would like to reduce down to a one pixel width (i.e. 1 y pixel per x value).

Thanks for any advice!

Chees

Answer 1

[mackhina]

Got it. Not the most elegant solution, but it works. Thanks for the tips!

size_image = size(image)
filt_image = zeros(size_image)
for j = 1:size_image(2)
    [rows, columns] = find((image(:,j)) > 0);
    row_index = round(mean(rows));
    filt_image(:,j) = image(:,j);
    filt_image(:,j) = 0;
    filt_image(row_index,j) = 1;
end

Answer 2

[Image Analyst]

Try bwmorph():

skeletonImage = bwmorph(binaryImage, 'skel', inf);

[mackhina]

This reduces the line down nearly all the way, but there are lots of instances where I will have multiple y values for a single value of x.

Answer 3

[Image Analyst]

Yes of course. Not every stretch will be a horizontal or diagonal stretch.

Answer 4

[Andrei Bobrov]

I = imread('line.png');
bw = im2double(rgb2gray(I));

[i,j] = find(bw);
[n,g] = findgroups(j);
idx = floor(splitapply(@median,i,n));

[k,l] = size(bw);
out = zeros([k,l]);
out(sub2ind([k,l],idx,g)) = 1;