Masters Theses

Abstract

"Early detection of malignant melanoma greatly benefits patients, as the overall success is dependent on finding these melanomas before they reach the invasive stage. Dermoscopy is a non-invasive skin imaging technique that studies have shown can improve the diagnostic accuracy of dermatologists by as much as 30% over clinical examination. In this project machine vision and image analysis techniques are used to detect annular granular areas in dermoscopy images automatically. The proposed algorithm utilizes the luminance ratio between annular and granular areas within the darkest 30% of the lesion. All points whose luminance value are less than 30% of the histogram are considered for further processing. The method has used some preprocessing steps to remove the unwanted effect of luminance reflection, to extract hair and bubble from the lesion image and to enhance the contrast of the image. Then the lesion plane is searched to find the center and border of annular-granular areas. Statistical analysis has shown that the implemented algorithm has the highest 92 percent in correct detection of annular granular areas"--Abstract, page iii.

Advisor(s)

Moss, Randy Hays, 1953-
Stoecker, William V.

Committee Member(s)

Stanley, R. Joe
Grant, Steven L.

Department(s)

Electrical and Computer Engineering

Degree Name

M.S. in Electrical Engineering

Publisher

Missouri University of Science and Technology

Publication Date

2012

Pagination

viii, 46 pages

Note about bibliography

Includes bibliographical references (pages 44-45).

Rights

© 2012 Parivash Hajiyani, All rights reserved.

Document Type

Thesis - Open Access

File Type

text

Language

English

Subject Headings

Melanoma -- Diagnosis
Skin -- Cancer -- Diagnosis
Image processing

Thesis Number

T 10552

Print OCLC #

903604830

Electronic OCLC #

903649673

Share

 
COinS