Image processing and image analysis in Bio7 using ImageJ, Python and R

Bio7 is an open source program that is built on top of the Eclipse rich client platform. Bio7 version 2.8 has a graphical user interface (R perspective) for the editing of code written in the programming language R and the EclipseImageJ1Plugin (ImageJ2 version 1.52) tightly integrated and simplifies the extraction of data from images, the data manipulation in a spreadsheet view using the programming language R and the conversion from data to image (plotting using R and imageJ) easy. It also has the plug-in GEF integrated which provides a flow chart editor that is used for the creation of documented and reproducible work flows and image/data processing pipelines. After installation of Bio7 and Python3 to the system partition (with the option “Add Python 3.x to PATH” during the installation selected), on computers with Microsoft Windows as the operating system, the environment variables might have to be edited to add the folder containing the python.exe to the system Path. Additional packages/libraries for Python provide easy access to important functions. Most python packages can be downloaded and installed via the command-line tool PIP that comes with the python3.x distribution for Windows from www.python.org. Ensure that PIP is properly installed. Edit the environment variables to add the folder …/Python3x/Scripts that contains the executable file pip.exe to the system Path. In order to download and install a python package that is available as source archives or as a python wheel in the python package index, start the command prompt (cmd.exe) and type the following into the command prompt and confirm by pressing the <enter> key

pip install <package>

or:

python -m pip install <package>

On Windows you can use the python launcher to specify a version:

py -3.7 -m pip install <package>

Here, <package> is a placeholder that has to be replaced with the name (without the <>) of the package on the python package index that you want to install. For example, to install the package NumPy:

py -m pip install numpy

The following packages are recommended: NumPy, DatabaseInterfaces, mysqlclient, Cython, Py2exe, Matplotlib, PyX, RPy, sh, Pillow, Pandas, pyserial, setuptools

Specify the path to to the executable file py.exe and python3X.exe. Then open Bio7 and in the main menu of Bio7 click Help ->install new software or Help -> Eclipse marketplace. Install the Eclipse plug-in PyDev. Then, in the main menu of Bio7 click Preferences -> preferences -> PyDev -> Interpreters -> Python Interpreter -> New. Enter the name to be displayed (e.g. “Python36”) and specify the path to the python36.exe. The Python interpreter menu has a tab “Packages”, where one can click on “Install/uninstall with PIP” in order to install new packages from within Bio7.

Image processing in Bio7 using the ImageJ perspective

In order to automate image processing, it is advisable to write imageJ macros or Scripts. Possibly the easiest approach is to write macros using the ImageJ macro language., create a New project (New -> Project) and give it some meaningful name. Add a new imageJ macro file to that project (New -> ImageJ -> ImageJ macro file (.IJM)) and call it appropriately. The final macro file has to be copied into the directory …\Bio7\plugins\com.eco.bio7.image_x.xx.xxx\macros. Finally, inside Bio7, open the ImageJ Edit perspective and click on the toolbar Plugins -> Macros -> Install and select the needed macro file. One macro file can contain several macros. One can also edit the file …\Bio7\plugins\com.eco.bio7.image_x.xx.xxx\macros\StartupMacros.txt and add custom macros to that macro file. After installing that file, those macros will be directly available after starting Bio7 from the ImageJ Edit perspective toolbar Plugins -> Macros without previously reinstalling the file using Plugins -> Macros -> Install in the ImageJ Edit perspective. I added the following code to the end of the file …\Bio7\plugins\com.eco.bio7.image_x.xx.xxx\macros\StartupMacros.txt:

// @title: ImageProcessingByDerAndere
// @author: DerAndere
// @created: 2018
// Copyright 2018 DerAndere
// @license: MIT
// @version: 0.0.1
// @language: ImageJ macro language
// @info: https://it-by-derandere.blogspot.com
// @description: ImageJ macros for image processing.

macro "Batch-AdustBCsetMaxValue" {

// This macro processes all the images in a folder and any subfolders.

  extension = ".tif";
  extension2 = ".czi";
  inputDir = getDirectory("Choose Source Directory ");
  channel = getNumber("which channel should be adjusted?", 0);
  maxVal = getNumber("maxVal for the chosen channel?", 65535);

  //dir2 = getDirectory("Choose Destination Directory ");
  outputDir = inputDir + "BCSetMaxValueAdjusted\\" + "C" + channel +"\\";
  if(!File.exists(outputDir)) {
    File.makeDirectory(outputDir);
  }
  setBatchMode(true);
  n = 0;
  processFolder(inputDir);

  function processFolder(inputDir) {
     list = getFileList(inputDir);
     for (i=0; i<list.length; i++) {
          if (endsWith(list[i], "/")) {
              processFolder(inputDir+list[i]);
          }
          else if (endsWith(list[i], extension) || endsWith(list[i], extension2)) {
             processImage(inputDir, list[i]);
          }
      }
  }

  function processImage(inputDir, name) {
       img = inputDir + name;
  // only proceeds if file is there
    if(File.exists(img)){
      // opens the images and gives them the temporary name "imgTemp" 
      open(img);
      ID = getImageID(img);
      selectImage(ID);
      dotIndex = indexOf(name, ".");
      basename = substring(name, 0, dotIndex);
      rename("imgTemp");
      // the next lines do your procedure
      Stack.setChannel(channel);
      stackSize = nSlices(imgTemp); // or nSlices(); or nSlices; ?
      for(z=1; z<=stackSize; ++z){
        setSlice(z) //?
        run("setMinAndMax", "min=0 max=maxVal");
        run("Apply LUT");
        wait(500);
        selectImage(ID);
        wait(500);
      }
    titleMod = "C" + channel + "adj";
    titleTitle = basename + titleMod + ".tif";
    wait(500);
    // add code here to analyze or process the image
    saveAs("tiff", outputDir+titleTitle);
    wait(500);
    close();
    }
  }
}

macro "SaveCurrentSliceAsTIF" {
id = getImageID();
title = getTitle(); // or getTitle;?
dotIndex = indexOf(title, ".");
basename = substring(title, 0, dotIndex);
path = getDirectory("image");
outputDir = inputDir + "Slices\\";
if(!File.exists(outputDir)) {
  File.makeDirectory(outputDir);
}
currentSliceNumber = getSliceNumber();
tileTitle = basename + "_S" + currentSliceNumber + "_mod1.tif";
// using the ampersand allows spaces in the tileTitle to be handled correctly
run("Duplicate...", "title=&tileTitle");
selectWindow(tileTitle);
saveAs("tiff",outpurDir+tileTitle);
close();
selectImage(id);
}
// End of ImageProcessingByDerAndere

Python scrips for image processing in the ImageJ perspective of Bio7

aaa

Transferring images into data using Bio7

aaa

Manipulation and analysis of data using the R perspective of Bio7

In order to use R from within Bio7, left-click R -> Preferences -> Preferences R. Select the correct Server. For Germany, try Germany (Goettingen). Left-click R-shell, enter update.packages() to update Rserve. If it is not installed, enter install.packages(“Rserve”). Execute by pressing enter. After the successful installation is confirmed in the R-shell, reboot your computer and restart Bio7 and left-click R-shell. Then, left-click the tool “Start Rserve” from the Bio7 toolbar at the top (or: R -> Start Rserve). Open the R perspective and Left-click “Packages” -> Install packages and search for packages you need. I recommend the packages dplyr and its dependencies that are installed automatically with it (dbplyr, RMySQL, DBI, lubrify, knitr, rmarkdown, ggplot2, RPostgreSQL, colorspace, utf8 and others), XLConnect, xlsx, foreign, tidyr.

Alternatives to Image J: Darktable, Krita, Gimp 3.2 or Endrov

Darktable is an excellent choice for radiometrically correct image processing. It allows you to completely deactivate any modules you don’t want to use, so you don’t even have to look at them, and also allows you to store presets for the remaining modules. darktable indicates the active modules with an icon that can be clicked to disable the module in question.

If you use Krita, open the image and left-click Left-click Image -> Properties… -> Image color space. Change Model to greyscale/alpha, bit depth (precision) 32bit floating point per channel. For 16 bit images in a color space with a linear tone reproduction curve (gamma = 1.0), select a Gray profile with linear tone reproduction curve (gamma = 1.0). An example is the ICC profile “Gray-D50-elle-V4-g10.icc” (note: The profile “Gray built-in (standard)” is the same as “Gray-D50-elle-V2-g22.icc” and has a tone reproduction curve with gamma = 2.2). Choose the rendering method “relative colorimetric”, activate “use black-point compensation” and left-click OK.. Image -> convert Image color space. Add a Layer on top of it and use it as an adjustment layer for non-destructive editing by selecting that layer with a left-click in the layer overview to make it the active layer and do level adjustments and/or inversion. By default, the final aggregate will be in the same color space as the bottom image. After level adjustments and inversions have been applied, change back to 16bit integer precision and the profile “Gray (built-in)” or better Gray-D50-elle-V2-srgbtrc.icc.

Gimp 3.2 (in development) will be fully color-managed.

Endrov is an alternative to ImageJ. Currently the contrast and brightness settings are too limited, but an ImageJ plug-in is available.

Icy includes ImageJ and Microscopy Manager

Deducer: a graphical user interface for the programming language R

Deducer is a specialized open source graphical user interface for data analysis that is built around the JGR console for the programming language R. In order to start Deducer from within the Bundled R installation that comes with Bio7, you have to locate the path for your Bio7 installation and open …/Bio7_x.x/plugins/Bundled_R_x.x.x/R/bin/x64 (all x stand for the version numbers installed). In that folder, create a shortcut to the file Rgui.exe and move this shortcut to a more convenient location. Open the Rgui by double clicking the freshly created shortcut. Install the packages JGR, Deducer and DeducerExtras, DeducerText. Under Microsoft Windows this can be done by selecting from the Rgui menu Packages -> Install packages…. Select a CTRAN mirror from the list and then select the packages JGR, Deducer and DeducerExtras. It is recommendet to install the following additional packages: DeducerSpatial, DeducerSurvival, DeduverText, ggplot2, Rcpp, RcppEigen, Rserve, tm, utf8, XLConnect, XLConnectJars. Click OK and wait until all packages are installed.

To start Deducer using the Rgui console, select from the Rgui menu Packages -> Load package…, select JGR from the list and click OK. Then type JGR() into the Rgui console and press <enter> to start the JGR console. After all required packages were loaded, Deducer’s Data Viewer opens. By hovering the mouse over the main menu of the JGR console (either File, Edit, Packages, Window or Help), the menu entries Data, Plot and DeducerExtra are automatically added to the JGR console.

In the JGR console, select from the menu Packages -> Package manager and choose the packages installed in the steps above to be loaded by default with the JGR console and click refresh. Read the documentation for R (here and here), for Deducer and for the packages mentioned above.

DERANDERE

Image processing