Comparative analysis of optimization algorithms for the reconstruction of all 3 elastic moduli in skin from noncontact OCE measurements utilizing Rayleigh waves

Many soft tissues (skin, cornea, muscle, tendon, etc.) are mechanically anisotropic and can be locally characterized with a nearly-incompressible transverse-isotropic (NITI) model. The phase velocity of Rayleigh wave propagating at an arbitrary angle to the symmetry (collagen fiber) axis is defined by all 3 elastic moduli of a NITI medium. However, reconstructing these moduli from measured Rayleigh-wave angular anisotropy is not trivial. This study compares simplex, genetic and deep-learning-based reconstruction algorithms to minimize time and improve accuracy for noisy experimental data. The algorithms were evaluated through simulations with varying noise levels and experimentally validated using dynamic OCE.