Statistical analysis to develop a three-dimensional surface model of a midsize-male foot

Abstract

A representative midsize-male foot was generated via a statistical analysis of foot scans from 107 men with widely varying body size. Seventy-two surface landmarks were manually extracted from the original scan data. A template fitting method was used to represent each scan with a homologous mesh. A principal component analysis and least-squares linear regression were used to generate a foot surface model with landmarks using a reference stature of 1755 mm and a body mass of 83.19 kg. The statistical model can be used to generate a wide range of male foot sizes and shapes.