morphmet  

Re: Morphometrics using R

morphmet
Fri, 11 Dec 2009 12:50:38 -0800



-------- Original Message --------
Subject: Re: Morphometrics using R
Date: Fri, 11 Dec 2009 19:23:08 GMT
From: Joseph Kunkel <j...@bio.umass.edu>
To: morphmet@morphometrics.org
CC: Joseph Kunkel <j...@bio.umass.edu>
References: <4b2288b4.4060...@morphometrics.org>

With respect to the image file formats readable by R it should be
noted that R has been a little conservative on reading files that are
lossy or of ill defined content such as TIFFs and indexed images such
as PNGs but has provided ways of printing output of images to such
files for a long time.  While it is nice to now be able to read such
files into R for useful and decorative reasons the younger or
beginning R user should know the evils of the TIFF, PNG and JPEG
formats from a quantitative perspective.

The TIFF file format produced by various softwares does not always
conform to a standard that can be read by a good Graphic Image
Manipulation Program.  Some hardware companies produce image TIFFs
that can only be read by their own software based on the ease of
implementing your own custom format according to the TIFF standard.

The JPEG is often used as a lossy (compressed to the point of
irreversible data loss) file for compactness but each time it is saved
it may lose more resolution unless the user is careful not to modify it.

The PNG file is indexed with a color palette and although it may
contain all the image detail that you want, the numbers that are used
to index the palette are likely meaningless as data.

The preferred image input for R has been the less well known PNM, PPM,
PGM, PBM file formats which provide non-lossy binary, grey scale or
RGB images that can be save in ASCII or binary file format.

These formats are good for data that you want to be able to read and
want to be rational in their construction.  You could read these files
with a word processor if you had a mind to.

They have the benefit of being non-lossy in data as well as having the
benefit of being a standard format that many software packages can
read (GIMP, R) and manipulate unlike plain ill defined raster images
that do not have dimension or layer information well specified.

If landmarks are one's main objective and the image is of sufficient
non-degraded resolution then by all means use the TIFF, PNG and JPEG
formats.  But be mindful of the reasons that R has been slow to
develop libraries to read these formats.

Joe Kunkel


On Dec 11, 2009, at 7:00 PM, morphmet wrote:



-------- Original Message --------
Subject: Re: Morphometrics using R
Date: Fri, 11 Dec 2009 17:41:56 GMT
From: Kevin Middleton <k...@csusb.edu>
To: morphmet@morphometrics.org
References: <4b210421.2080...@morphometrics.org>

I am attempting to learn how to do morphometrics on R using the book from Julien Claude and to visualize the jpegs I need to install the package rimage. However, the package will not install ("error: C++ compiler cannot create executables") and it automatically uninstalls the package. It is quoted as being "ORPHANED" on the website. Is there a comparable package for jpegs on R that anyone knows of? Thanks!

Both the biOps (http://cran.r-project.org/web/packages/biOps/index.html ) and ReadImages (http://cran.r-project.org/web/packages/ReadImages/index.html ) packages have jpeg reading functions.


Kevin


-------------------------------------------------
Kevin M. Middleton
Department of Biology
California State University San Bernardino




--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org

-·.  .· ·.  .><((((º>·.  .· ·.  .><((((º>·.  .· ·.  .><((((º> .··.·
=-       =º}}}}}><
Joseph G. Kunkel, Professor
Biology Department
University of Massachusetts Amherst
Amherst MA 01003
http://www.bio.umass.edu/biology/kunkel/





--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org