3D Optical Body Composition Analyzer
A web-based tool implementing a peer-reviewed algorithm for automated body-composition analysis from 3D optical scans. Accepts mesh-file uploads and returns templated meshes with body-composition predictions for demonstration purposes.
Reference publication
Tian IY, Wong MC, Kennedy SF, Kelly NN, Liu YE, Garber AK, Heymsfield SB, Curless B, Shepherd JA. A Device-Agnostic Shape Model for Automated Body Composition Estimates from 3D Optical Scans. Medical Physics 2022.
Abstract
Background. Many predictors of morbidity caused by metabolic disease are associated with body shape. 3D optical (3DO) scanning captures body shape and has been shown to accurately and precisely predict body-composition variables associated with mortality risk — with advantages in safety, cost, and accessibility compared to dual-energy X-ray absorptiometry (DXA). However, standardization across manufacturers remains lacking.
Purpose. We introduce a scanner-agnostic algorithm that automatically fits a topologically consistent human mesh to 3DO-scanned point clouds and predicts clinically important body metrics using standardized models.
Methods. Training used 848 scans across three different 3D optical systems (ages 18–89; BMI 14–52 kg/m²). The pipeline applied template-mesh registration, anatomically constrained PCA deformation, and surface-to-surface optimization. A unified PCA model generated body-composition predictions, learned from 562 withheld test scans.
Results. The model achieved R² > 0.8 for most predictions (highest 0.94 for total fat / lean and trunk fat). Root-mean-squared errors stayed below 3.0 kg. Repeatability precision was 2–3× worse than DXA, except for visceral fat measurements.
Conclusions. The method provides an accurate, automated, and scanner-agnostic framework for standardizing 3DO scans — a radiation-free, low-cost alternative to criterion radiology imaging for body-composition analysis.