Body shape and size modelling using regression analysis and neural network prediction

The aim of this research is to build a patient-specific virtual body shape model for patients of various Body Mass Index (BMI) and body shape. This will enable simulated epidural procedure on patients of various body characteristics, to increase trainee skill, reduce injuries and litigation costs. Regression analysis (RA) and artificial neural networks (ANN) were implemented to accurately calculate body shape in a data-driven approach. Epidural simulator software was developed containing a screen to enter patient characteristics. When the patient BMI is adjusted, the modelled body shape and tissue layer thickness updates allowing patient specific simulation. The model uses anthropometric measurements as input: body mass, height, age, gender and body shape. The developed model enables a virtual representation of any actual patient to be built based on their measured parameters for epidural rehearsal prior to in-vivo procedure. Copyright © 2014 by ASME.