CorrelationFunctions.jl is a library to compute all classical correlation
functions from 2D and 3D images.
Currently, we provide functions to compute:
-
$S_2$ (two-point probability) -
$L_2$ (lineal path function) -
$C_2$ (cluster function) -
$F_{ss}$ (surface-surface function) -
$F_{sv}$ (surface-void function) -
$P$ (pore-size function) -
$p$ (chord length function) -
$\rho$ (phase cross-correlation function) -
$F_{sss}$ (surface-surface-surface function) -
$F_{svv}$ ,$F_{ssv}$ , … (other three-point surface-void functions)
Some additional three-point or non-classical CFs currently under
development. The general scheme of computations for all major functions is shown
in figure below:
Salvatore Torquato's book "Random Heterogeneous Materials" provides a comprehensive theoretical summary on all classical CFs.
The package is available through Julia's Pkg ecosystem.
For example, from Julia REPL: import Pkg; Pkg.add("CorrelationFunctions")
All functions are described in the documentation, you can also get help on each
function in the REPL (using ?). The documentation for the most recent release
is available here on
GitHub Pages.
Alternatively, to build a documentation locally do the following:
- From Julia REPL:
import Pkg; Pkg.add("Documenter") - From shell, this directory being the working directory:
cd docs && julia make.jl
Numerous Jupiter notebooks with examples of how to apply the package to compute various correlation functions are available in our FaT iMP research group's repository, e.g.:
- Example1: general
- Example2: surface functions
- Example3: additional surface functions
- Example4: 3-point correlation functions
Video lecture is available here
The functionality of the package is described in following scientific papers, please, support us by citations if you find our code useful:
- The main paper describing the package and its CPU and GPU implementations for all classical CFs Postnicov, V., Samarin, A., Karsanina, M. V., Gravey M., Khlyupin, A. & Gerke, K. M. (2023). Evaluation of classical correlation functions from 2/3D images on CPU and GPU architectures: introducing CorrelationFunctions.jl. Computer Physics Communications, 299, 109134.
- The digital approach to compute 2-point surface functions; also this paper introduces the C0.5 criterion (the package contains the function to evaluate it for input images) to judge the quality of the image to access such CFs Samarin, A., Postnicov, V., Karsanina, M. V., Lavrukhin, E. V., Gafurova, D., Evstigneev, N. M., Khlyupin, A. & Gerke, K. M. (2023). Robust surface-correlation-function evaluation from experimental discrete digital images. Physical Review E, 107(6), 065306.
- The truly exact approach to compute 2- and 3-point correlation functions on smooth boundary sets, improved edge filter, computation of 3-point surface functions for digital images Postnicov, V., Karsanina, M. V., Khlyupin, A., & Gerke, K. M. (2023). The 2-and 3-point surface correlation functions calculations: From novel exact continuous approach to improving methodology for discrete images. Physica A: Statistical Mechanics and its Applications, 628, 129137.
- Evaluation of 3-point correlation functions from structural images on CPU and GPU architectures: accounting for anisotropy effects [Postnicov, V., Karsanina, M.V., Khlyupin, A., Gerke, K.M. (2024) in revision with Physical Review E.]