Image Manipulation, with an Emphasis on Journaling

Standard imaging operations, e.g. crop and contrast adjustment, but with ability to go back and forth through sequence of changes, with records being persistent. Optional Shiny interface. Useful to help with the research reproducibility problem, and as a teaching tool.


Imaging package, with an emphasis on journaling, i.e. recording history of changes. Undo/redo operations, ability to display multiple versions (currently under construction), etc. The history is persistent, i.e. across sessions. Can be run from the R command line, or from a Shiny-based GUI.

Installation

You will need the following packages for the command-line interface to the package:

  • fftwtools: Install from CRAN, except for Linux; for the latter, sse these special instructions.
  • EBImage: Run these commands from within R:

    source("http://bioconductor.org/biocLite.R", verbose = FALSE) #Install package
    biocLite("EBImage", suppressUpdates=TRUE, suppressAutoUpdate=FALSE, ask = FALSE)
  • GUI Installation: Install from CRAN.
    install.packages(c('shiny','shinyjs'))

Having done this, you can install ShinyImage. For instance, download the .zip package available here and unpack it, creating a directory/folder ShinyImage-master. Then from a terminal window, run

R CMD build ShinyImage-master
R CMD INSTALL -l z ShinyImage_0.1.0.tar.gz

with z being the location you wish to install ShinyImg to (changing the version number as necessary).

Alternatively, ShinyImage can be installed using devtools. User's working directory must be set to ShinyImage-master. From R,

> devtools::install()

Example Usage

Here we will perform several actions, both to illustrate some ShinyImage operations and also to show the journaling. All operations will use the R command line; examples of the GUI are given later in this document.

# load image, whether local file or from the Web
# the image being used is titled 'A tiger in the water'
#  By Bob Jagendorf 
#  [CC BY 2.0 (http://creativecommons.org/licenses/by/2.0)], 
#  via Wikimedia Commons
> tiger <- 
   shinyimg$new("https://upload.wikimedia.org/wikipedia/commons/1/1c/Tigerwater_edit2.jpg")
 
# 'tiger' is an object of class 'shinyimg', which in turn is a subclass
# of 'R6'
 
# set autodisplay on 
# after first image modification, 
# image will render and pop up in a new window
# or user can manually render image
> tiger$set_autodisplay()
# manually rendering image
> tiger$render()
# crop the image
> tiger$crop()
[1] "Select the two opposite corners of a rectangle on the plot."
# add brightness
> tiger$add_brightness()
# add contrast
> tiger$add_contrast()
# add gamma
> tiger$add_gamma()
# add blur
> tiger$add_blur()
 
 
# remove brightness
> tiger$remove_brightness()
# remove contrast
> tiger$remove_contrast()
# remove gamma
> tiger$remove_gamma()
# remove blur
> tiger$remove_blur()
 
# we have had nine actions, and can undo the last 8 of them
# we will undo the last five actions (remove blur and remove gamma)
# by calling undo five times
# undoes the removal of the blur
> tiger$undo()
# undoes the removal of gamma 
> tiger$undo()
# undoes the removal of contrast
> tiger$undo()
#undoes the removal of brightness
> tiger$undo()
# undoes the adding of the blur 
> tiger$undo()
# we can also redo the adding of the blur
> tiger$redo()
 
# we can also save the image to edit later on
> tiger$save("tiger-water.si")
# and later we can come back after a cold boot to do:
> tiger <- shinyload("tiger-water.si")
# if you want to revert to a previous saved state, you can also do:
> tiger$load("tiger-water.si")
# this will load the image back to the state it was in when you saved the image.
> tiger$undo()  # not too late to undo changes made before the save!
 
# lastly, if we want to save the physical image
> tiger$saveImage('tiger.jpg')
# we can save it as either jpg, png, or tiff
> tiger$saveImage('tiger.png')
# if a user does not specify the name, the default is temp.jpeg
> tiger$saveImage()

GUI Installation and Usage

Installation

Download from CRAN:

install.packages(c('shiny','shinyjs'))

Usage

  • A gui can also be spawned to edit images using the sample provided or a user can upload an image, link, or an image log of a .si object created through shinyimg

  • A user can edit brightness, contrast, and gamma correction. The user can also rotate, blur, and crop an image. These changes can be made to the image using the sliders. In order to crop a photo, the user has to highlight a box over the original plot. A preview of the cropped image with pop up below the original image. To keep the cropped image, click the keep button which will pop up below the preview image.

  • While editing an image, a user can undo, redo, or reset the image. These actions are executed through buttons at the bottom of the sidebar.

  • After editing an image, a user can download the image and the image log. These actions are below the main plot.

  • The user can also view the image log to see which actions were recorded.

    Run these commands from within R.

    > runShiny()
  • A user can use an image they are currently editing on the commandline to edit in the GUI

    # using our previous example of our shinyimg object tiger
    > runShiny(tiger)

    Installing fftwtools on Linux

    Installing fftwtools on Linux

    • Download http://www.fftw.org/fftw-3.3.6-pl1.tar.gz
    • Unpack, say to x/fftw-3.3.6-pl1 and from that directory run

      ./configure --prefix=y --enable-shared=yes 
      

      where y is your desired installation directory for fftwtools, say /usr/local.

    • Run the usual make; make install sequence.
    • Set environment variables (no spaces around the = sign!):

      export C_INCLUDE_PATH=x/fftw-3.3.6-pl1/api 
      export LD_RUN_PATH=y/lib 
      export LIBRARY_PATH=y/lib 
      
    • You may need to install libtiff-dev, say by

      sudo apt-get install libtiff-dev
      
    • You may also need to install fftw-dev, by

      sudo apt-get install fftw-dev
      
    • If fftw library does not get properly installed, try

      sudo apt-get install fftw3 fftw3-dev pkg-config
      
    • Then run the R steps as above.
  • News

    Reference manual

    It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

    install.packages("ShinyImage")

    0.1.0 by Norm Matloff, a year ago


    Browse source code at https://github.com/cran/ShinyImage


    Authors: Alexander Fu [aut] , Ariel Shin [aut] , Norm Matloff [aut, cre]


    Documentation:   PDF Manual  


    MIT + file LICENSE license


    Depends on EBImage, R6, shiny, shinyjs

    Suggests knitr, rmarkdown


    See at CRAN