I want to remove some artifacts from skin lesion images, then compute the hair detection performance; but I don't know howto get the hair ground truth? Can I do the hair tracking manually instead of by expert?
If you are working with color images, then work on the first image cnannel (luminance channel);
a=imread('image.tif");
b=a(:,:,1);
Then apply some filtering on the image b. Try to do some median filtering, which is suitable for additive noise as also proposed by Joachim. This will eventually, by repeating filtering, remove the artifacts.
Then in order to find the objects try first with thresholding using the im2bw function.
Then you may also try clustering or active contours.
It depends what you want to do. If you want to identify the hair then maybe you need to do texture analysisis. The hair has a different texture that the rest of the image.
Please read some theory about texture and segmentation:
See
C.P. Loizou, C. Theofanous, M. Pantziaris, T. Kasparis, “Despeckle filtering software toolbox for ultrasound imaging of the common carotid artery”, Comput. Meth. & Progr. Biomed., vol. 114, pp. 109-124, 2014. Best paper, editor’s choice.
C.P. Loizou, C.S. Pattichis, “Despeckle filtering algorithms and software for ultrasound imaging,” Synthesis Lectures on Algorithms and Software for Engineering, Ed. Morgan & Claypool Publishers, 1537 Fourth Street, Suite 228, San Rafael, CA 94901 USA, June 2008, ISBN-13: 9781598296204.
- Badges
- Science topic