Monte Carlo simulations of the Radiative Transport Equation are very popular in the field of biomedical optics (and beyond).
There has recently been a good deal of interest in coherent scattering effects, such as the "memory effect". All monte carlo models that I am aware of are unable to account for coherent effects in a physically consistent manner because scattering events occur non-deterministically through out the sample.
This repository contains a coherent monte carlo simulator that is able to properly model coherent scattering effects.
The key insight is that the scattering events must be deterministic. This is accomplished by keeping track of all the individual scatter locations. Keeping track of the individual scatterer positions ensures that the scattering, while random, it is deterministic. This is phyiscally more accurate---as we all know, speckle patterns generated from scattering through a static object, while random, are deterministic.
There are many other monte carlo models that are capable of producing speckle patterns. Speckle patterns are the most visible coherent effect, however other more subtle coherent effects such as the memory effect are not so easily described. Furthermore, most such simulations, if run a second time (or even if the number of photons in the simulation is altered) will produce a different speckle pattern.
Assumptions behind the model:
- Uses a discrete scatter approximation
- The sample is a single slab
- The edges of the slab have periodic boundary conditions
- There is no absorption
- Scatters are isotropic
The core algorithm for the simulation is written in c (using the c99 standard). The simulation is written as a library, with the interface described in coherentmc.h. The library is compiled into a shared object, which can be accessed however you like. All of the core c-code is in the coherentmc subdirectory.
I provide a nice python wrapper (that contains more details about the underlying c functions) that uses ctypes to access the library. The library is in the root of the repo. The library abstracts away the calling process into two objects and two functions. There is a Setup object, that describes the setup, and there is a Result object that contains the results of a simulation run. There are two functions, a run function, and a run_again. function.
Some utility functions for working with the cmc module are contained in the cmcutils module.