how to do the distance transform

Hi,all.I am processing a image. I need to do the distance transformation.This is my code:

myImage=imread('EarImage_1.tif');
I = rgb2gray(myImage);
whitePart = I > 220;
blackPart = I < 110;
imshow(whitePart);
I2=1-blackPart;
figure,imshow(I2);
D=bwdist(I2);
figure,imshow(~D);

This is my original picture:

After this process,I got an image just the same as "I2",which is the binary image before. I didn't get the picture as expected. I am a beginner of matlab. If anyone could tell me the solution of that, I would appreciate very much.

Thank you.

[Walter Roberson]

http://www.mathworks.com/matlabcentral/answers/7924-where-can-i-upload-images-and-files-for-use-on-matlab-answers

Answer 1

[Sean de Wolski]

showing ~D means that only distances equal to zero (i.e. points that are true in I2) will be shown. D is an array of doubles.

imshow(D,[]);

will show you the distance xform image with the full range of values. Good job getting to where you are, +1.

[Yu Zhang]

Thank you very much! I got it. But the results look not good. I could not identify the kernel at all. Could you tell me whether I should do some other processes before the distance transform? I have done the median filter and get the gradient of the image. The gradient image looks good which is clear enough.Thank you!

[Sean de Wolski]

We need to see your image. Post it somewhere, perhaps to one of the websites in Walter's comment link and paste the url into your original question.

Answer 2

[Walter Roberson]

Do not subtract blackPart from 1 if you want to see white pixels for the locations selected by blackPart.

1-(I<110) is the same as (I>=110) which could look very similar to (I>220) that you are already plotting. All of the (I>220) would be included within the (I>=110)