Optimizing Monte Carlo simulation for detecting the internal information in a fat-muscle media

摘要

To verify the optimized Monte Carlo simulation for obtaining the best source-detector separation to acquire the inner information in turbid media, the spatially diffusive spectral system, which took fat-muscle tissue as example, was build. According to the typical banana-shape visiting probability profile, increasing the restricted conditions was applied to simulate a spatial filter, which was used to weaken the influence of the overlying layer and reject multiply scattered photons. First, study the relationship between the effective signal (I_S(z_2,r_i), the non-effective signal (I_M(z_1,r_i) + I_N(z_2,r_i)),the effective signal ratio (SNR) and the source-detector separations (SDS) when fat thickness varied from 0.1 to 0.55 cm. Secondly, study the relationship between h_f and SDS_(best). Simulation results showed the optimized MC simulation, which can gain more information than original MC simulation, can be used for detecting the internal information in multilayered tissue, and SNR can be improved, and h_f is used as the independent variable to develop a linear regression model to predict SDS_(best). (R~2 = 0.9808). The method is expected to provide more evidence for quick disease check-up in vivo and is instructive for select the best source-detector separation.