We will discuss the importance of scattering phase function models and implementation of the probe-sample boundary for accurate Monte Carlo (MC) simulations of reflectance. Deep learningaided regression models will be considered for improving the efficiency of MC simulations. We will briefly introduce our recently published Python library with a massively parallel OpenCL implementation of the MC method (https://github.com/xopto/pyxopto) for layered and voxelized media. Liquid optical phantoms based on monodisperse spherical particles will be discussed for connecting the MC simulated and the corresponding measured reflectance through calibration. Finally, we will discuss deep learning-based regression models for real-time estimation of optical properties.

LU ASI Zinātniskajā seminārā savu prezentāciju par tēmu "Optical Fiber Probes: From Monte Carlo Simulations to Estimation of Optical Properties" demonstrēs Prof. Miran Burmen no Ļubļanas Universitātes.