bwboundaries: Explanation of a example

This simple example is given in the Matlab Documentation:

I = imread('rice.png');
BW = im2bw(I, graythresh(I));
B = bwboundaries(BW,'noholes');
for k = 1:length(B)
    boundary = B{k};
end

Even if I can see the result and why we need the loop, I don't fully get to understand what exactly is doing the "for" loop, I wouldn't know how to explain how it works. Can someone help me? Thank you

Answer 1

[Image Analyst]

If you have 10 blobs, you would get 10 boundaries. It gets the k'th boundary with this line

boundary = B{k};

And then it gets the x and y coordinates. Think of it like this

for k = 1 : length(B)
     thisBoundary = B{k};  % Get k'th boundary
     x = thisBoundary(:, 2);
     y = thisBoundary(:, 1);
     plot(x, y, 'w', 'LineWidth', 2);
end

The 'w' means to plot in white color, and the line width is the thickness of the line it's using to plot the boundary in the overlay above the image.

[Zynk]

Thanks, I understood why I needed the loop in order to plot,but not exactly how it was getting the boundary or why. Now it is clear.

Answer 2

[Anand]

I assume you are talking about Example 1 in the documentation.

         I = imread('rice.png');
         BW = im2bw(I, graythresh(I));
         [B,L] = bwboundaries(BW,'noholes');
         imshow(label2rgb(L, @jet, [.5 .5 .5]))
         hold on
         for k = 1:length(B)
             boundary = B{k};
             plot(boundary(:,2), boundary(:,1), 'w', 'LineWidth', 2)
         end

The for loop here is used in order to plot a white outline displaying the boundaries on the image.

[MashalS]

what does noholes mean? what if its replaced by holes in 3rd line ?

[Image Analyst]

noholes means that there are no nested boundaries. So a donut shaped region would return only the larger outer boundary, not the inner boundary between the donut and the hole.

if you did not use noholes, for a donut region, it will return both the outer boundary and the inner boundary between the region and the hole. In other words, it will return the boundary of the hole (same as inner boundary of the donut) as well as the outer boundary of the donut.

[Suresh D]

             plot(boundary(:,2), boundary(:,1), 'w', 'LineWidth', 2)
Please explain this line

[Image Analyst]

boundary is an n-by-2 array of the row, column coordinates, in other words

[row1, col1;

row2, col2;

...

rown, coln];

which is

[y1, x1;

y2, x2;

...

yn, xn]

so boundary(:,2) is a column vector of all the x values and boundary(:,1) is a column vector of all the y values. So the plot command is essentially

plot(x, y, 'w', 'LineWidth', 2);

which will draw a white line between all the boundary points.

Answer 3

[Babu Sankhi]

For my images I got the boundaries only for the very good contrast png images not for the less contrast images. Is there any way to get boundary even for the images with less contrast? In addition to that I want to find the distance between center and boundary(left/right) for several images. Is there any helpful codes/function so that I can find those distances more precisely?

Thank you

[Image Analyst]

Depends on the image. You can try adapthisteq().

[Babu Sankhi]

Thank you analyst,

I am still gettting difficulty on findig the codes so that I can measure the distance between rightmost edges of green and red boundaries ( I mean the distance between A and B points) of the attached png image? Can You please help me on this regard? I have several images to do that .

Thank you

[Image Analyst]

You can use bwdist(). Attach your original image in a new question if you still need help.

[Babu Sankhi]

Thank you analyst for your help. But bwdist() did not give any difinite value in my case so I am still stucked with that . I have attached my original microscopy image and the code I used to get the boudaries of desired region of image . After that my problem is to find the distance between right most points of innner(green) and outer(Red) boundaries . It means I have to find the distance between A and B in in resulting png.

Can you please share more ideas about it ?

thank you