基于神经网络法反求耳结构弹性模量

Inverse derivative of elastic modulus for human ear based on neural network

目的通过耳结构的位移反算耳结构弹性模量。方法基于Patran软件建立耳结构有限元模型,使用Mat-lab建立计算耳结构反问题的BP神经网络。对耳结构有限元模型进行频率响应分析,得到鼓膜凸和镫骨足板的位移响应;把位移作为BP神经网络的输入、相对应的结构弹性模量作为输出,对网络进行训练。结果利用训练成熟的BP网络反算出耳结构的弹性模量,相对误差非常小。结论反算结果表明,所使用的反问题方法求解耳结构弹性模量是可行的,可为临床提供确定生物结构力学参数简捷有效的方法。

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 re sponse 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.