Research On Biological Tissue Optical Model By Monte Carlo Simulation

Accurate measurement of tissue optical parameters is a very important field of biological and medical applications, and it is an important prerequisite for optical basic research and clinical applications in the biomedical field. There are two problems in this area. Problem1is how to use mathematical methods to describe photons’ transmission through biological tissue and problem2is how to measure tissue optical parameters accurately. The two problems are not solved completely. Monte Carlo simulation is an effective way to describe photons’ transmission in biological tissue. This research uses CONV program written by Wang Lihong and CUDAMCML software written by Erik Alerstam to simulate photons’ transmission in layered tissue model, and gets the data of luminous flux and diffuse reflectance. Luminous flux and diffuse reflectance are used as inputs, and optical parameters are used as outputs. Establish BP Neural Network, and train this network to inverse the tissue optical parameters, such as scattering coefficient and absorption coefficient. The main contents of this research involve:(1) Introduce Monte Carlo Simulation Method to research the process that photons are transmitting through the tissue.(2) Using Monte Carlo Simulation Program to build single-layer model, two-layer model and three-layer model. Change single-layer, two-layer and three-layer models’ optical parameters, and simulate these models by Monte Carlo method. And then get the data of diffuse reflectance, and analysis the relationship of the parameters and diffuse reflectance. Plot the optical image of the tissue by Matlab software. Get the relationship of the parameters and luminous flux through optical image.(3) Using BP Neural Network, luminous flux or diffuse reflectance is as inputs, and scattering coefficient and absorption coefficient are as outputs. Simulate single-layer, two-layer and three-layer models by Monet Carlo program.Through the results, get training and verify databases. Use training database to train the BP Neural Network. Use verify database to verify the accuracy of the net. By the neural network, the scattering coefficient and absorption coefficient can be inversed.(4) Based on BP Neural Network, use GUIDE from Matlab to design optical parameters inversing system which is a visual interface. Import luminous flux, scattering coefficient and absorption coefficient database to this system, and it can predict scattering coefficient and absorption coefficient accurately.By error analysis, the error of this BP network is relatively small, and the prediction accuracy is high. Optical parameters inversing system can provide reference information for clinical diagnostic and treatment, so it has practical significance.