Title

Modified Watershed Technique and Post-Processing for Segmentation of Skin Lesions in Dermoscopy Images

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.

Department(s)

Electrical and Computer Engineering

Second Department

Chemistry

Sponsor(s)

National Institute of Health (U.S.)

Keywords and Phrases

Border detection; Dermoscopy images; Ground truth; Malignant melanoma; Neural network classifier; Noise removal; Post processing; Segmentation; Skin lesion; Watershed segmentation; Dermatology; Imaging systems; Landforms; Neural networks; Oncology; Watersheds; Image segmentation; artificial neural network; benign skin tumor; epiluminescence microscopy; false positive result; hair; image processing; melanoma; noise; priority journal; skin tumor; vignette; Algorithms; Dermoscopy; Humans; Image Enhancement; Image Interpretation; Computer-Assisted; Pattern Recognition; Automated; Reproducibility of Results; Sensitivity and Specificity; Skin Neoplasms; Neural network; Watershed

Library of Congress Subject Headings

Image processing

International Standard Serial Number (ISSN)

0895-6111

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2011 Elsevier, All rights reserved.

Share

 
COinS