1Department of Scientific Computing, Florida State University, 2US Army Natick Soldier Research, Development and Engineering Center
Saturday All day, Clinch Concourse
Characterization of morphometric variation within a targeted population is a critical component in the design of personal protective equipment and the establishment of standards for its effective fitting. In this poster, we report on the initial evaluation of a large set of facial landmark data for a sample of U.S. Marines. Generalized Procrustes analysis removed variation from the data introduced by location, scale, and orientation. The effect of size was reintroduced, leaving only information about form. A principal component analysis revealed patterns of covariation in the data. The first three principal component axes captured 21%, 14%, and 8% of the total variation respectively, while 21 axes were required to capture 90%. Visualization of the first two principal component axes suggested that the first accounts for face size and the second accounts for face height verses width; results that are consistent with previous research into facial variation related to respirator fit testing. Visualization of the third principal component axis accounts for interpupillary breadth relative to face size. This analysis is the first step in developing models that will exploit these patterns in the prediction of overall head shape from craniofacial landmarks and ultimately provide the tools for exploiting much larger, but incomplete, data sets that could be used to inform the design, manufacture, and fitting of protective and therapeutic head gear in military, industrial, and medical fields.
This work was funded, in part, by Cooperative Research Agreement W911QY-12-2-0004, between Florida State University and the US Army Natick Soldier Research, Development, and Engineering Center.