Date of Graduation

2015

Document Type

Thesis

Degree Type

MS

College

Eberly College of Arts and Sciences

Department

Forensic and Investigative Science

Committee Chair

Thirimachos Bourlai

Committee Co-Chair

Patrick Buzzini

Committee Member

Tina Moroose

Abstract

In response to criticism of traditional forensic methods of human head hair examination, the goal of this study was to develop and evaluate an objective method using texture-based image analysis. 120 individuals, including 20 from six different maternal groups, were sampled and their hairs evaluated using the developed method. Different variables were evaluated to develop the most efficient method to increase intra/inter ratio, as well as the percentage of well-classified hairs. The variables included different microscope systems, objective lenses, number of hairs examined per individual, hair shaft regions, focus methods, and normalization filter techniques. Statistical analyses included one-way analysis of variance, agglomerative hierarchical clustering, and classification trees. A tested method based on the combinations tested and a theoretical method based on evaluation of the performance of each variable were determined. The tested method revealed increased ratio of inter- and intra-variability and percentage of well-classified hairs when the reference population is decreased. The method also revealed percentages of well-classified hairs above 90% when comparing an individual to other members of the same maternal group, a comparison that would fail to differentiate if tested with mitochondrial DNA analysis. Future testing of the theoretical method, as well as further testing of maternal groups and population size, will need to be done to verify these results.

Download

Share

COinS