second derivative of an image

Hi every body

The contour of an object in one image is known as the points with second derivative equal to zero. How can i get this characteristic knowing that I have only the image (the pixels intensity)??

Please can any one help me, thanks a lot

Answer 1

[Nitin]

There is a nice explanation on image derivative here . Please check it out, slide 23 onwards.

[jan]

Thanks Nitin for your answer but what i really need is the second derivative :\

[Nitin]

Did you check the Laplacian second derivative on slide 31?

[jan]

Hi Nitin In fact, to obtain this characteristic we can apply the laplacien filter and this is what i tested before asking my question and this is the code that i used:

    I=imread('image.png');
    w=fspecial('laplacian',0); 
    g=imfilter(I,w,'replicate');
    figure(1),imshow(g,[]);

but in g, the contour points aren't only the points with derivative equal to zeros there are other points inside and outside the object with derivative equal to zeros. So what i really need what i have tested is it true? and any explication or other suggested ideas because i need to characterise, in a formal manner the contour points to do a clustering thanks

[Image Analyst]

Why don't you post your image to help with your explanation about what you want to measure in the image?