Algorithm to Automatically Extract Body Sizes and Shapes

Abstract

This study presents an algorithm to automatically extract the size and body shape of a 3D scanned model. The methods used in this research include factor analysis, linear regression equation, cluster analysis, and discriminant analysis. These are used to analyze the body’s shape and choose the best primary dimensions for establishing the sizing system table. Authors use fuzzy logic to establish the mathematical model. In this model, the input variables are the inseam height and the neck girth measurements, and the output variables are the numbers of the human size coding and body shape. In addition, the rotation matrix and the optimal function are used to write an algorithm to estimate the neck girth and inseam measurements. Furthermore, a simple approach based on vertices and surface normal vector data, together with optimal searching, is adapted to estimate the primary dimensions. This estimation algorithm, combined with the fuzzy logic model, makes the automated process of extracting the size and body shape possible. The findings of the study suggest a new research method for quickly informing people about their body shape. This supports purchasing clothes and designing tailored clothing. The automatic algorithm will be very useful for buying clothes face-to-face or online.