®®®® SIIA Público

Título del libro: International Conference On Artificial Intelligence, Computer, Data Sciences, And Applications, Acdsa 2026
Título del capítulo: 3D Body Silhouette Modeling System for Morphological Analysis: Proof of Concept

Autores UNAM:
ALFONSO GASTELUM STROZZI; CELIA ANGELINA SANCHEZ PEREZ;
Autores externos:

Idioma:

Año de publicación:
2026
Palabras clave:

3-D vision; Body fat prediction; Body fats; Body volume; Digital neck circumference; Modelling systems; Morphological analysis; Proof of concept; Relative errors; Volumetrics


Resumen:

© 2026 IEEE.Accurate, accessible, and non-invasive anthropometric assessment is essential for monitoring healthrelated morphological changes in clinical and wellness applications. This work presents a 3D vision system based on the Intel RealSense D455 camera, capable of capturing depth maps to obtain digital measurements of neck circumference. This measurement is used as an input variable for two computational approaches: (i) predictive models trained on a public database to estimate body fat percentage and body weight, and (ii) a volumetric analysis method that quantifies excess volume based on a body silhouette tracing model that compares a volumegained silhouette with an initial reference silhouette. These models are evaluated through a proof-of-concept study to demonstrate their usefulness as quantitative monitoring tools. The system achieved cervical circumference measurements with relative errors below 1%. The regression models exhibited adequate performance in weight prediction (R2=0.84) and provided a consistent estimation of body fat. In addition, the volumetric analysis detected small volume increases with relative errors below 6%. The proposed system offers a low-cost, noninvasive alternative for digital anthropometric measurement, estimation of body-related indicators, and detection of volumetric changes, supporting morphometric analysis applications in health and wellness contexts.


Entidades citadas de la UNAM: