Steve on Image ProcessingJuly 14th, 2006Seen on the web: MATLAB example for tracing and measuring skin lesionsI saw an interesting MATLAB based image processing application on a melanoma skin cancer blog recently. The example, including code, shows how the lesions are segmented, traced, and measured. Several Image Processing Toolbox functions are used. Here's a thumbnail version of a screen-shot from the blog:
Check it out! 2 Responses to “Seen on the web: MATLAB example for tracing and measuring skin lesions” |
|
|||
|
These postings are the author's and don't necessarily represent the opinions of The MathWorks. |
Nice presentation of how to trace the skin lesions. I have a
similar problem. In my case I try to trace the boundaries of
different nuclei.
I have to major cases: isolated nuclei and overlapping nuclei. I managed to get a black and white image and my question now is how to distinguish when there is a single nuclei or two overlapping nuclei. One idea is that a single nuclei is a convex area and two overlapping nuclei is a concave area. Is there any method that can tell me if a given labelled object is convex or concave? Due to irregularities, also for the single nuclei, some parts of the contour might look concave.
Br,
Andrea
Andrea—You can use the ‘Image’ and ‘ConvexImage’ properties returned by regionprops. You can use a logical operation to find the object pixels in “concave” regions. Then maybe you can look at the ratio of the number of pixels in concave regions to the total number of object pixels.