Masters Theses

Keywords and Phrases

Melanoma; Seborrheic Keratosis; Dermoscopy; Inflamed Keratin Plugs; Adaptive Thresholding; Absolute Thresholding; Histogram; Feature Extraction; Classifier


“Malignant melanoma is a very deadly form of skin cancer which has claimed many lives over the past few years. If detected early this can be cured, hence early detection of malignant melanoma is essential. Unfortunately melanoma is mimicked by seborrheic keratosis, a benign skin cancer. Identifying malignant melanoma as seborrheic keratosis using clinical diagnosis can prove fatal to the patient. To prevent such errors, dermoscopy, a common non-invasive skin imaging technique, is used which improves the diagnosis of these pigmented lesions by visualizing the morphological structures. This study proposes an automatic method by applying image processing techniques to aid in dermoscopy. The purpose of this study is to differentiate melanoma from seborrheic keratosis by applying thresholding techniques to the dermoscopy images. The algorithm consists of absolute thresholding of the red chromaticity plane and adaptive thresholding of the green and blue planes to detect inflamed keratin plugs in the images. The parameters for thresholding are obtained from histogram analysis. The images obtained after applying this technique are then processed to extract different features such as color and texture features. The information obtained from the feature extraction is given to a classifier to differentiate melanoma from seborrheic keratosis. The proposed algorithm is applied on a dataset consisting of 369 melanomas and 256 seborrheic keratoses. This method yielded 94.0% accuracy with 98.6% of melanomas correctly identified”--Abstract, page iii.


Moss, Randy Hays, 1953-

Committee Member(s)

Stoecker, William V.
Stanley, R. Joe
Shrestha, Bijaya


Electrical and Computer Engineering

Degree Name

M.S. in Electrical Engineering


Missouri University of Science and Technology

Publication Date

Summer 2016


viii, 42 pages

Note about bibliography

Includes bibliographic references (pages 40-41).


© 2016 Pramada Kishtagari, All rights reserved.

Document Type

Thesis - Open Access

File Type




Thesis Number

T 12024

Electronic OCLC #