PhenoLink

Description

PhenoLink, former "Checkout" is a multipurpose open source high content image analysis software developped at Ksilink entirely written in C++ for computational speed purposes.
Dedicated to biological images produced with automated microscopes, Checkout has a ability to display and analyze all types of images (2D, 3D, 2D+t, 3D+t, Multichannels, etc.) in a speedy and efficient manner due to different level of multi-processing granularities.

Features

• Multi Readers

  Mainly developed for Yokogawa (CV7000/CV8000), Checkout also handles InCell and Zeiss readers. More readers can be of course implemented.

• Image Processing

OPENCV
DLIB
Cell Profiler

• Data Mining

• Deep Learning

• Visualization

• Scability

• Data Management

Getting Started

Installation

First you need a fully working C++ compiler suite (On windows Visual Studio 2017 up to 2022 where tested and ubuntu 18.04 / 20.04 with gcc)

We use Vcpkg for easier package integration, this builds all the needed dependencies for you

** cmake **

VCPKG based installation:

        git clone https://github.com/Microsoft/vcpkg.git  
        # if you have issue with vcpkg updates we tested with commit 0a3bc3d25 
        # git checkout 0a3bc3d25 # successfull vcpkg we are able to build against

On Windows: .\vcpkg\bootstrap-vcpkg.bat

.\bootstrap-vcpkg.bat .\vcpkg.exe integrate install .\vcpkg.exe upgrade --no-dry-run # if you want to update existing vcpkg #

Some libraries are needed to get started usually downloaded automatically during the installation process.

Dependencies

• OPENCV : for all image processing
• DLIB : specific library (Images)
• MongoDB : Metadata management
• ONNX : Deep Learning inference (Conversion Python model to Onnx)
• ARROW (Apache) : Data efficient and fast storage (feather output file, pandas:pd.read_feather())
• Qt : GUI
• WNDCharm : Multiple C++ basic image processing features (similar to cellpainting features)
• Boost : WndCharm needs it.

Optional

• VTK : 3D Visualization
• Python3 : Data Post processing, analysis and vizualisation
• CellProfiler : If cellpainting protocol necessary, PhenoLink handles all the data mapping for you and can launch a Pipeline of CellProfiler within it's plugin framework

From Sources

• Windows

  git clone https://github.com/Ksilink/PhenoLink.git      
  
  mkdir  PhenoLinkBuild
  cd PhenoLinkBuild
  
  
  cmake ..\PhenoLink -DCMAKE_TOOLCHAIN_FILE=%Path_to_vcpkg%\vcpkg\scripts\buildsystems\vcpkg.cmake  -DVCPKG_TARGET_TRIPLET=x64-windows -G"Ninja Multi-Config" -DVCPKG_INSTALL_OPTIONS=--x-buildtrees-root=d:/bt -DCMAKE_BUILD_TYPE=Release
  

  or

  cmake ..\PhenoLink -DCMAKE_TOOLCHAIN_FILE=%Path_to_vcpkg%\vcpkg\scripts\buildsystems\vcpkg.cmake  -DVCPKG_TARGET_TRIPLET=x64-windows -G"Visual Studio 17 2022" -DVCPKG_INSTALL_OPTIONS=--x-buildtrees-root=d:/bt -DCMAKE_BUILD_TYPE=Release
  

Then perform the build:

  cmake  --build . -j 12 --config Release

You can pass --clean-first to the build command line to be sure to have a clear new build

• Linux

  export Qt6_DIR= %Path_to_QT%/Qt/6.X.X/gcc_64/
  
  git clone https://github.com/Ksilink/PhenoLink.git      
  
  
  cd PhenoLink && git pull
  
  
  cd ../PhenoLink
  
  
  cmake ../PhenoLink/ -DCMAKE_TOOLCHAIN_FILE=${HOME}/Code/vcpkg/scripts/buildsystems/vcpkg.cmake -DCMAKE_BUILD_TYPE=Release || exit 1
  
  cmake  --build . -j 72 --config Release --clean-first || exit 1
  

GPU Version

if you want to build with the ONNX library using GPU add to the cmake configuration command line:

  -DONNX_GPU:BOOL=ON 

Releases

A release versuib for Windows:x64 can be downloaded from here : PhenoLink