Inverse derivative of elastic modulus for human ear based on neural network / 医用生物力学

ABSTRACT

Objective To obtain elastic modulus through displacement of the ear structure. Methods The finite element model (FEM) of human ear structure based on Patran software was constructed and the neural network for inverse derivative of elastic modulus for the ear was established using Matlab software. The frequency response of the ear structure FEM was calculated to obtain the displacements of tympanic membrane and stapes. The displacements acting as input data of training samples and the corresponding elastic modulus acting as output data were used to train the neural network. Results The elastic modulus was inversely derived by adopting this mature neural network with relatively less error. Conclusions The viability of the proposed methods for inverse derivative of elastic modulus was demonstrated in this paper, which could provide a simple and effective method to obtain mechanical parameters for clinic work.