have a data of melanoma images,I need to extract the following features from them : 1-Asymmetry feature 2-Color variation feature 3-Border feature 4-Diameter feature To use them in the ABCD rule for melanoma detection

asked 2016-06-28 19:54:18 -0500

Maria Ateeq gravatar image

updated 2016-06-29 07:56:48 -0500

hey i want help for C++ opencv coding.

I have a data of melanoma images, I segmented them now I need to extract the following ABCD features from them to detect melanoma skin cancer and these ABCD features are as follow:

Asymmetry.One half doesn't match the appearance of the other half.

Border irregularity. The edges are ragged, notched, or blurred.border is not in uniform shape.

Color. The color (pigmentation) is not uniform. Shades of tan, brown, and black are present. Dashes of red, white, and blue add to a mottled appearance.

Diameter.*The size of the mole is greater than 1/4 inch (6 mm), about the size of a pencil eraser. Any growth of a mole should be evaluated.

Could anyone help me to do this job please? Or have a C++ opencv code to calculate these features? Thanks in advance.

edit retag flag offensive close merge delete


unfortunately , you're assuming that anyone here knows what is meant by "Border-feature".

i guess, you'll have to explain all of those, to get any help.

berak gravatar imageberak ( 2016-06-28 20:51:49 -0500 )edit

@berak yes you are right , I am adding more details now .Thankyou

Maria Ateeq gravatar imageMaria Ateeq ( 2016-06-29 07:51:12 -0500 )edit

while you're waiting, bookmark the docs page (you'll need the search box later..)

for sure you will need to find contours, so have a look at the tutorials (the contourArea() is already half the solution to D, i guess)

berak gravatar imageberak ( 2016-06-29 08:32:33 -0500 )edit

btw, cv::meanStdDev() on some small roi might solve problem C.

if you could add an example image ?

berak gravatar imageberak ( 2016-06-29 21:13:59 -0500 )edit

For the asymmetry and border irregularity part, take a look at convex defects searching using OpenCV. It has been done for hands, but I guess it should work on moles also.

StevenPuttemans gravatar imageStevenPuttemans ( 2016-06-30 06:32:09 -0500 )edit

BTW, this problem is solved at almost every single computer vision/medical conference I go to in the last 2 years, so don't reinvent the weel and take a look at the available research on the internet!

StevenPuttemans gravatar imageStevenPuttemans ( 2016-06-30 06:34:16 -0500 )edit

Did you make it? If yes, plis post, will help me a lot! Thanks

ggrayce gravatar imageggrayce ( 2018-08-05 00:07:09 -0500 )edit

@ggrayce, there is no reason in reviving a topic from 2016, especially since the feedback to the author was, go have a look around to what is already out there :D Make your own new question if you need help on this.

StevenPuttemans gravatar imageStevenPuttemans ( 2018-08-06 07:48:00 -0500 )edit