Automated Melanoma Recognition

 The goal of this project is to find an Automated Melanoma Recognition algorithm based on dermoscopy images. Below, you can see outline of the project.

	Preliminaries       Because of the higher incident of malignant melanoma, researchers are more and more concerned with automated diagnosis of skin lesions. Although melanoma is not the most frequent cancer, it is the fastest growing cancer in the world. For more information you can see www.dermoscopy.org .
	Skin Lesion Segmentation       In almost every vision project, segmentation is a bottleneck to reach. Since, we have reached over 90 percent of correct segmentation. To see the results click here. Software package will be available in February 2006.
	Skin Lesion Feature Extraction       This part of the project is under research now.
	Feature Classification       This part of the project is under research now.

• Preliminaries

• Skin Lesion Segmentation Results

                Some of our segmentation results are in the table below. You can see that segmentation of pigmented lesions is not affected by hairs or    

          skin texture and color. Both the bright and dark skin lesions are correctly segmented. Multiple spots are considered.

                To see more results click here.

• Skin Lesion Feature Extraction

• Feature Classification

• Related links

Author: Nima Razavi.
© Copyright 2003-2005 - Institute for Studies in Theoretical Physics and Mathematics (IPM).
All rights reserved. Please submit your comments or questions here.
Last update:10/13/2005.