In order to obtain the number of pixels of the yellow line which you know it's real distance is 11.72 mm in the vertebral column image, you could try this
% Draw a line convering the desired distance and get the number of pixels
imshow(my_image)
% Retrieve the intensity values of pixels along the line for 1000 points
[cx,cy, rgbValues, xi,yi] = improfile(1000);
oimg = findobj(gca,'Type','image'); %image which is open
img = get(oimg, 'CData'); % image data
d_line = round(sqrt((xi(1)-xi(2))^2 + (yi(1)-yi(2))^2))
[cx,cy, rgbValues] = improfile(img, xi, yi, d_line);
rgbValues = squeeze(rgbValues); % change dim
% distance in pixels of the line drawn
d_pixels= sqrt( (xi(2)-xi(1)).^2 + (yi(2)-yi(1)).^2);
% plot drawn line
hold on;
plot(xi, yi, 'r-');
For the first image you would finally know the distance in pixels and the real distance (mm) , being able to define the conversion factor: mm/px or px/mm, but only for elements in that specific image.
In order to compare between the pixel distances of each image to see if they match, you should get a scale factor of each image (haven't tried the code right now)
scaleFactor1=d_pixels/image1_rows
And for the second image the same,
scaleFactor2=height_px/image2_rows
Then you could scale one image to the other and finally compare the heights.
S=scaleFactor1/scaleFactor2
height_px_scaled=S*height_px
