Format

Send to

Choose Destination
Comput Med Imaging Graph. 2011 Mar;35(2):116-20. doi: 10.1016/j.compmedimag.2010.09.006. Epub 2010 Oct 20.

Modified watershed technique and post-processing for segmentation of skin lesions in dermoscopy images.

Author information

1
Department of Electrical and Computer Engineering, Missouri University of Science and Technology, Rolla, MO, USA. hwqkc@mst.edu

Abstract

In previous research, a watershed-based algorithm was shown to be useful for automatic lesion segmentation in dermoscopy images, and was tested on a set of 100 benign and malignant melanoma images with the average of three sets of dermatologist-drawn borders used as the ground truth, resulting in an overall error of 15.98%. In this study, to reduce the border detection errors, a neural network classifier was utilized to improve the first-pass watershed segmentation; a novel "edge object value (EOV) threshold" method was used to remove large light blobs near the lesion boundary; and a noise removal procedure was applied to reduce the peninsula-shaped false-positive areas. As a result, an overall error of 11.09% was achieved.

PMID:
20970307
PMCID:
PMC3183575
DOI:
10.1016/j.compmedimag.2010.09.006
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Elsevier Science Icon for PubMed Central
Loading ...
Support Center