Yu-Chin Chai

Graduation Semester and Year

2006

Language

English

Document Type

Thesis

Degree Name

Master of Science in Computer Science

Department

Computer Science and Engineering

First Advisor

Manfred Huber

Abstract

Fatal skin lesions such as melanomas appear as having high color variation to experts but often as having low variation to novices. This thesis investigated if it is possible to find a function representing an expert's color variation sensitivity to skin lesions without explicitly requiring them to provide a diagnosis. Various pigmented skin images with either normal pigmentation or with diseased lesions were evaluated by experts and novices. To determine the best color space for lesion differentiation, an optimization process based on parametric color dimensions in a RGB color space was used to find the parameters of a color transformation function that best differentiated the color variation sensitivity between experts' and novices'. The analysis of results showed that this process was able to maximize the differences between the expert's and novice's color variation sensitivity and to provide a criterion to discriminate between images of non-invasion and melanoma type skin lesions.

Disciplines

Computer Sciences | Physical Sciences and Mathematics

Comments

Degree granted by The University of Texas at Arlington

Download

Share

COinS