Detect the boundary in a 2 clustered colored image

I have this image. It has 2 clustered areas, a yellow and a blue one. I also drew 2 centrized points. I want to to find out which point is closer to the boundary which is why I want the boundary in the first place. I have to note that I am aware of the coordinates of the points but I can't use a specific starting point (for example in order to use the command bwboundary, because I have about 5000 images and don't know which one the function rand() will choose each time). All of them share the same philosophy.

Answer 1

[Image Analyst]

You get a mask for the yellow, then call bwboundaries and then use sqrt (optional) and min:

boundaries = bwboundaries(yellowMask);
b = boundaries{1};
xb = b(:, 2);
yb = b(:, 1);
% Determine distances from point 1.
distances1 = sqrt((x1 - xb) .^ 2 + (y1 - yb) .^ 2);
minDistance1 = min(distances1)
% Determine distances from point 2.
distances2 = sqrt((x2 - xb) .^ 2 + (y2 - yb) .^ 2);
minDistance2 = min(distances2)

Knowing that you can find out if point 1 or point 2 is closer to a boundary point.