| Summary: | The aim of this thesis was to conduct an anthropometric survey of the New Zealand Defence Force (NZDF) and to explore potential implications and applications of 3D anthropometric data in a military clothing context. Anthropometric data are traditionally captured by trained measurers using basic tools such as tape measures and callipers. However, these measurements can be time consuming and limited in their ability to capture true human body size and shape. The emergence of 3D body scanning can reduce measurement time while increasing the volume of measurements collected. Anthropometric datasets are beneficial to the military in terms of health, design, ergonomics, engineering, and in particular, clothing applications. Although numerous countries have conducted anthropometric surveys, research on the application of 3D anthropometric data for military clothing is limited. The New Zealand Defence Force had never had an official 3D anthropometric dataset but expressed an interest in its application to their clothing processes. This thesis is structured into eight sequential chapters dedicated to developing and delivering a quality 3D survey methodology and protocol (Chapters 2-4), implementing the survey and building the 3D database (Chapter 5), and exploring the applications of this database to the military (Chapters 6-8). The dataset collected in Chapter 5 consisted of traditional and 3D anthropometric measurements for 1,003 (791 male: 212 female) NZDF uniformed personnel. The final scan database contained 3,009 body scan images and 84,252 body measurements. Furthermore, the study produced a comprehensive protocol consisting of measurement and landmarking instructions for all 84 measurements collected in the survey, in addition to their respective summary statistics. Chapter 6 identified whether there was sexual dimorphism in the NZDF by examining the absolute and proportional differences in anthropometric measurements between males and females. The findings showed that males had significantly greater absolute and proportional measurements compared to females for most measurements, except waist and thigh girth. Additionally, a combination of waist girth, hip girth, bicep girth, and biacromial breadth, could classify sex with 95% accuracy. These results could be used to guide the design of male and female variants for clothing and other military equipment. The purpose of Chapter 7 was to determine how well decision tree models could predict tailor-assigned combat uniform sizes (the current unisex shirt and trouser in the NZ Army), and whether automatic, post-processed, or physical anthropometry measurements led to the most accurate model. The findings showed that classification was more accurate when based on the 3D automated measurements and was higher for trousers (62.8% accuracy) compared to shirts (54.9%). Body weight and waist girth were identified as the most important variables in the decision tree models. Despite the modest accuracy, these findings demonstrate that 3D anthropometric measurements, when used with decision tree models, show promise for classifying unisex clothing sizes. If refined, this system could increase the efficiency of the uniform sizing process for future military recruits. Chapter 8 attempted to elucidate the modest prediction results in Chapter 7 by determining how anthropometric characteristics could be clustered into distinct groups and examining the characteristics of each cluster. Principal Component Analysis was used to identify key body dimensions for upper (waist girth and body height) and lower body (waist girth, hip girth, and inseam length) clustering. A combination of two-step and k-means cluster analysis resulted in six and 10 distinct clusters for female and males, respectively. Body scan images representing each cluster centroid were presented to show the variation in body shapes. The resulting cluster information could be used to develop gender-specific sizing systems. This body of work demonstrates how traditional and 3D body scanning measurements can be used in a comprehensive anthropometric survey, measurement protocol, and database for the NZDF. This thesis illustrates that 3D anthropometric data can be applied successfully to the military. More specifically, 3D anthropometric databases have been shown to be useful for (1) identification of sex-related differences within a military population, (2) prediction of military clothing sizes using automated measurements, and (3) identification of body size groups or 'clusters' that can be used to develop sex-specific sizing systems. It is hoped the original information contained within this thesis will help military organisations to conduct large-scale 3D anthropometric surveys and to develop 3D anthropometric databases, which, in turn, can help provide solutions to real-world clothing-, ergonomics-, engineering-, and health-related problems.
|
|---|
