Use of Body Shape Information in Clothing Size Selection

Abstract

To investigate the role of body shape information on clothing size selection, a sample of 143 males were measured and sized using a computerized digital-image based measurement system. Clothing sizes were initially determined by the system using traditional criteria for the long sleeve shirt, jacket and trousers of a military dress uniform. The best-fitting size was determined by trial and error based on subjective feedback and expert judgement, provided by clothing and sizing technicians. Discriminant function analysis was used to determine sizing rules for each garment, based on different sets of anthropometric input variables. Comparisons were made between the prediction performances of discriminant functions derived from traditional variables and those of functions derived from 3D landmark coordinates. The results indicate that the use of three-dimensional landmark coordinates, as input to a discriminant function analysis, is superior to the use of circumference measurements in predicting clothing sizes. The use of these landmarks is thought to improve the classification of cases by allowing a better characterization of body shape.