A Bayesian approach to accounting for variability in mechanical properties in biomaterials

TL;DR

This paper introduces a Bayesian framework for modeling variability in biomaterial properties, integrating experimental data and hyperelastic models to improve characterization and classification of bio-tissues.

Contribution

It presents a novel Bayesian approach that treats model parameters as random variables, enabling probabilistic characterization and classification of biomaterials.

Findings

Successfully characterized sheep arteries using Bayesian inference

Developed a Bayesian classification method for biomaterial data

Demonstrated the approach's effectiveness with experimental data

Abstract

In this paper, we present an approach for modeling bio-tissues that incorporates the variability in properties as part of their characteristics. This is achieved by considering the parameters of the model of a biomaterial to themselves be random variables and represented by a probability distribution over the space of parameters. This probability distribution is obtained by the systematic use of Bayesian inference together with a continuum mechanics based solution of a boundary value problem. We illustrate this approach by characterizing sheep arteries by using a combination of experimental data and different hyperelastic models. Furthermore, we also develop a model based Bayesian classification of new data into different classes based on the computed model parameter probability distribution.