Differences

This shows you the differences between two versions of the page.

--- documents:101105ij_r_jython [2014/04/14 21:57] – [2. Using ImageJ from R] kota
+++ documents:101105ij_r_jython [2020/11/26 09:11] (current) – external edit 127.0.0.1
@@ Line 1: / Line 1: @@
+======= R + ImageJ =======
+Analysis using several tools is a powerful way to deal with complex Image data. We often do this, and here is one small example of taking advantages of ImageJ and R at the same time. There are three ways:
+  - Use R from ImageJ (Rserve)
+  - Use ImageJ from R (no description yet)
+  - Use ImageJ and R from Jython REPL
+====== 1. Using R from ImageJ via Rserve ======
+In this example, R functions are used from ImageJ via Rserve using Jython scripting.
+[[http://www.rforge.net/Rserve/doc.html|Rserve]] is a package that you could install from within R. In addition, you need two jar files in the plugin folder of ImageJ/Fiji.
+  * [[http://www.rforge.net/Rserve/files/|Rserve downloads]]
+After the installation, Rserve is an independent program so you could directly start it from the commandline (Rserve is a server that waits for commands, exectutes them and return the output) by:
+<code>
+R CMD Rserve
+</code>
+Then in the Fiji script editor, try the following code:
+<sxh python>
+from org.rosuda.REngine.Rserve import RConnection
+c = RConnection()
+x = c.eval("R.version.string")
+print x.asString()
+c.close()
+</sxh>
+If you see the version number of R printed in the output field, then you are successfully communicating with R from ImageJ.
+[[http://rforge.net/org/doc/|Here is the Javadoc of Rserve]].
+[[http://cmci.embl.de/documents/120206pyip_cooking/python_imagej_cookbook#rmulti-peak_fitting_using_r|Here is an example]] of using multi-peak fitting and shoing the peak positions within the current image.
+====== 2. Using ImageJ from R ======
+One easy way is to execute a shell command to run ImageJ macro / scripts in R. Such example is well documented in the link below.
+http://stackoverflow.com/questions/18989725/how-can-i-call-execute-an-imagej-macro-with-r
+I have not made any attempt to directly access ImageJ classes from R but here is my advice for an inquiry. There is an R-package "RImageJ" that allows you to use ImageJ commands. Problem is that this package has not been updated since 2010.
+<blockquote>
+ RImageJ:
+http://romainfrancois.blog.free.fr/index.php?category/R-package/RImageJ
+This basically allows use of methods available in one of the classes in ImageJ called ij.IJ
+http://rsbweb.nih.gov/ij/developer/api/ij/IJ.html
+This class IJ is a kind of utility class with many static methods that allows user to operate on images like ImageJ macro (especially the **run()** method does sort of such).
+Access is limited, but depending on what you want to do, I could maybe advice more.
+Another option is to access at much lower level using rJava:
+http://www.rforge.net/rJava/index.html
+By adding ImageJ to your Java path, it's probably possible to write script in R like Rhino javascript or Jython.
+</blockquote>
+====== 3. Using ImageJ and R from Jython Interactive Interpreter ======
+ImageJ classes could be accessed by Jython, a python based java interface. R has rJava (JRI, Java - R Interface). For this reason, one way of using ImageJ functions and R functions in an integrated environment is to use these resources from Jython.
+Python-scripting environment is light enough to test many things. After these testings are done, then a serious Java code could be written.
+Here is a trial in win32. All commands will be done from command line. It should basically be similar in other environment.
+===== Setting Up Environment. =====
+You should have done:
+  * [[http://rosuda.org/rJava/|rJava]] package should be installed in R.
+  * [[http://www.jython.org/|jython.jar]] should be somewhere locally.
+**1. set CLASSPATH**
+Following three jars should be in your classpath.
+  * JRI.jar
+  * REngine.jar
+  * JRIEngine.jar
+  * ij.jar
+... so in command line
+<code>
+set CLASSPATH=%CLASSPATH%;C:\Program Files\R\R-2.12.0\library\rJava\jri\JRI.jar;C:\Program Files\R\R-2.12.0\library\rJava\jri\REngine.jar;C:\Program Files\R\R-2.12.0\library\rJava\jri\JRIEngine.jar;C:\ImageJ2\ij.jar
+</code>
+**2.  set PATH**
+R native libraries should be linked. Then the path should be added with bin, jri, R.dll and JVM.dll:
+<code>
+set PATH=%PATH%;C:\Program Files\R\R-2.12.0\bin;C:\Program Files\R\R-2.12.0\library\rJava\jri;C:\Program Files\R\R-2.12.0\bin\i386;C:\Program Files\Java\jre6\bin\client
+</code>
+===== Example Scripting =====
+<sxh python>
+C:\>java org.python.util.jython
+Jython 2.5.1 (Release_2_5_1:6813, Sep 26 2009, 13:47:54)
+[Java HotSpot(TM) Client VM (Sun Microsystems Inc.)] on java1.6.0_22
+Type "help", "copyright", "credits" or "license" for more information.
+#importing image using ImageJ
+>>> import ij.IJ
+>>> import org.rosuda.JRI.REXP
+>>> import org.rosuda.JRI.Rengine as Rengine
+>>> import ij.IJ as IJ
+>>> imp = IJ.openImage("http://rsb.info.nih.gov/ij/images/blobs.gif")
+>>> imp.show()
+>>> flA = imp.getProcessor().getFloatArray()
+>>> intA = imp.getProcessor().getIntArray()
+>>> intA3=intA[3]
+>>> histA = imp.getProcessor().getHistogram()
+# going into R, first instantiate
+>>> engine = Rengine(['--no-save'], False, None)
+>>> engine.assign("histR", histA)
+>>> engine.eval("pixint <- {0:255}")
+>>> engine.eval("histform <- data.frame(pixint, histR)")
+>>> engine.eval("plot(histform$pixint, histform$histR)")
+#.. will plot the graph but one cannot close it...
+>>> engine.eval("graphics.off()")
+#will close all the plots.
+</sxh>