AnimalProcessingRoutines

Overview

To simplify the compilation of c++ routines used in the Matlab Animal Processing Tools we propose the python based approach using SCons project. Here we describe step by step the compilation processes. Note what we use ITK static library build, not dynamic.

Requirements

• Package Scons version > 4.4, see https://scons.org/pages/download.html

  • the simplest way to obtain Scons is using conda (see https://www.anaconda.com/)

    conda install -c conda-forge scons

• Visual Studio 2022, Visual Studio 2019 should do the job as well, see https://visualstudio.microsoft.com/downloads/

• you wil nedd c++ boost, because few routines has abusive implementation using boost

  • open Windows Power Shell (run as Administrator)

    Register-PackageSource -provider NuGet -name nugetRepository -location https://www.nuget.org/api/v2

    Install-Package boost_filesystem-vc143

    Install-Package boost_program_options-vc143

• download ITK https://itk.org/download/

• note ITK do not support scons compilation, so install CMake GUI https://cmake.org/download/

• if you need compile SegmentedImageAnalysis, a priori, Matlab packages should be installed first.

  • System Properties - Advanced - Environment Variables - Edit Environment Variables : add 'C:\Program Files\MATLAB\R2018b\bin\win64'

Project Compilation

ITK build

we will use STATIC library build

• open CMake GUI

  cmake-gui

  in CMake do :

  • indicate where is downloaded ITK derectorie in Where is the source code tab
  • indicate where will be your build in Where to build the binaries tab (usual practice is to create build derectorie inside the source code folder)
  • if build derectory does not exist, CMake will propose to create it
  • Now select Configure
    • specify the generator for the this project, when CMake ask you : Visual Studio 2019 or Visual Studio 2022
    • verify the options:
      • BUILD_EXAMPLES uncheck
      • BUILD_STATIC_LIBS option should be check
      • BUILD_TESTING uncheck
      • CMAKE_CONFIGURATION_TYPES = Release, leaves only Release
      • ITK_BUILD_DEFAULT_MODULES should be check
  • after configuration finished select Generate
  • when Open project, if CMake do not propose the rigth Porject Generator (Visual Studio > 2019):
    • find in build the file 'ITK.sln'
    • select 'Open with' and choose the rigth version of Visual Studio to build the project

• in Visual Studio, from ALL_BUILD select build and wait untill ITK installation finished

SConstruct file adoptation

in project directory AnimalProcessingRoutines edit the file SConstruct :

• give the path of itk library by setting the variable "itkPath"
• give the path of boost installation
• give the path of matlab installation
• check the libraries labels

Compile project with Scons

now we can compile the project, from command line in Windows PowerShell or Terminal

cls; scons

wait until Scons compile the project. The executable binaries will be stored in release directory following the tree :

└── release
    ├── ct
    |    └── cellTracking.exe
    ├── jt
    |    └── jonctionTracking.exe
    ├── sa
    |    └── segmentedImageAnalysis.exe
    ├── zg
    |    └── zmapGeneration.exe
    └── zp
         └── zmapProjection.exe