ImageJ - Fwd: interpolation Note of Caution

ImageJ

Fwd: interpolation Note of Caution

Posted by John Oreopoulos on May 12, 2008; 11:24pm
URL: http://imagej.273.s1.nabble.com/Fwd-interpolation-Note-of-Caution-tp3696252.html

There is a very interesting topic discussion regarding image
manipulation going on at the confocal microscopy listserver (see
forwarded thread discussion) and I was wondering if someone has
already created an ImageJ plugin that records what actions have been
performed on an image or an image stack.

John Oreopoulos

Begin forwarded message:

> From: Chris Tully <[hidden email]>
> Date: May 12, 2008 6:57:35 PM EDT (CA)
> To: [hidden email]
> Subject: Re: interpolation Note of Caution
> Reply-To: Confocal Microscopy List <[hidden email]>
>
> Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-
> bin/wa?S1=confocal List members,
>
> First let me explain that although I am no longer associated with
> them, I have previously worked for Media Cybernetics.
>
> One extremely valuable set of features in Image-Pro Plus
> (www.mediacy.com) is the Audit Trail and Image or File Signature.
> The Audit Trail logs every action taken on every open image. This
> allows you to document exactly what was done with or to the image.
> The Image and File Signatures are 32 bit check sums that are
> automatically recorded in the Audit trail at relevant times
> (Acquisition, save, load...), and are sensitive enough to detect a
> change of +/- one gray level in one pixel! Paired with the Capture
> module's Auto Save function, it is possible to:
>
> 1) Document that a published image is unchanged. You will need to
> carefully track such things as cropping to demonstrate this
> completely, but this is entirely possible.
>
> 2) If the image has been altered use successive image signatures
> (before and after each alteration) to demonstrate that the logged
> alterations are the only ones that have occurred. If you are going
> to do this I would recommend saving the image with a new name
> immediately before any such alteration so that you can demonstrate
> the alteration again if challenged.
>
> Another approach that I often use is to work on a duplicate of the
> original image. Change away as much as I need to to do the desired
> analysis. As the last step of my analysis though I generate
> outlines of the objects that I am measuring and place them back on
> the unaltered original image. This both allows me to make some
> measurements that can only be made on the original image and to
> demonstrate that the identified objects are still relevant to the
> original image and therefore to the sample.
>
> Chris
>
> --
> Chris Tully
> Microscopy and Image Analysis Expert
> [hidden email]
> 240-888-1021
> http://www.linkedin.com/in/christully
>
>
> On Mon, May 12, 2008 at 3:35 PM, RICHARD BURRY <[hidden email]>
> wrote:
> Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-
> bin/wa?S1=confocal
> The guidelines Doug posted on the Univ. Arizona website are great.
> One issue that makes digital images more suspect, is the individual
> element of the image, the pixel, can be changed. With photographic
> manipulations we were burning in regions not selectively changing
> the intensity of a single pixel. Most journals today reserve the
> right to ask authors for the original image files that were taken
> on the CCD camera or came from the confocal. As listed by Doug in
> his guidelines, it is most important to retain archive files of ALL
> original images. This allows you to go back to the original if
> needed.
>
> One mistake that authors are making in submitted manuscripts is to
> saturate the intense pixels in images. The images look almost like
> line drawings and not micrographs. It is important to keep all the
> information in the images when processing and to adjust the
> settings when collecting images to spread the intensity over the
> full range.
>
> Dick Burry
> Ohio State University
>
>
>
> ----- Original Message -----
> From: Doug Cromey <[hidden email]>
> Date: Monday, May 12, 2008 12:54 pm
> Subject: Re: interpolation Note of Caution
> To: [hidden email]
>
> > Search the CONFOCAL archive at
> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
> >
> > I've been thinking about the issue of digital imaging ethics for
> > awhile.Much of what we could get away with in the days of the
> > photographic darkroom
> > would no longer be considered appropriate these days. The
> > JCB has pretty
> > explicit digital image guidelines, I suspect that other journals
> > that are
> > without specific guidelines are probably "behind the
> > curve".
> >
> > My take has always been that if you fully describe the steps
> > that are taken
> > in processing an image, then no one can accuse you of misconduct
> (1).
> > Reviewers & Editors may not like your image processing protocol,
> > but then it
> > becomes an issue of scientific discussion, not an accusation.
> >
> > I've proposed some digital imaging ethical guidelines here:
> > http://swehsc.pharmacy.arizona.edu/exppath/micro/
> digimage_ethics.html
> >
> > Some colleagues at the University of Alabama - Birmingham are
> > working on a
> > web site that includes these guidelines and a video case study,
> > but it's not
> > quite done yet. I'll post the URL when the folks at UAB
> > let me know they
> > are done.
> >
> > Doug Cromey
> >
> >
> > (1) The HHS Office of Research Integrity officially defines
> scientific
> > misconduct as: ".fabrication, falsification, or plagiarism
> > in proposing,
> > performing, or reviewing research, or in reporting research
> results."
> >
> > * Fabrication is making up data or results and recording or
> > reporting them.
> > * Falsification is manipulating research materials, equipment,
> > or processes,
> > or changing or omitting data or results such that the research
> > is not
> > accurately represented in the research record.
> > * Plagiarism is the appropriation of another person's ideas,
> > processes,results, or words without giving appropriate credit.
> > * Research misconduct does not include differences of opinion.
> >
> > FROM: http://ori.hhs.gov/publications/ori_intro_text.shtml
> >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > Douglas W. Cromey, M.S. - Assistant Scientific Investigator
> > Dept. of Cell Biology & Anatomy, University of Arizona
> > 1501 N. Campbell Ave, Tucson, AZ 85724-5044 USA
> >
> > office: AHSC
> > 4212 email:
> > [hidden email]: 520-626-
>
> > 2824 fax: 520-626-2097
> >
> > http://swehsc.pharmacy.arizona.edu/exppath/
> > Home of: "Microscopy and Imaging Resources on the WWW"
> >
> >
> > -----Original Message-----
> > From: Confocal Microscopy List
> > [mailto:[hidden email]] On
> > Behalf Of MODEL, MICHAEL
> > Sent: Friday, May 09, 2008 7:43 AM
> > To: [hidden email]
> > Subject: Re: interpolation Note of Caution
> >
> > Search the CONFOCAL archive at
> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
> >
> > I certainly didn't' expect to start a big discussion of this
> > topic. I
> > agree that the Notes of Caution are fully justified. On the
> > other hand,
> > in Russ' Image Processing Handbook you will find many examples
> > of very
> > drastic editing of microscopic images, so in some situations it
> > must be
> > acceptable. Is it up to each journal to set up their own
> > guidelines? It
> > seems to me that so long as the author fully explained what had been
> > done to the images it's not cheating, but I may be wrong. (I
> > think we
> > already had this discussion on this forum before).
> >
> > -----Original Message-----
> > From: Confocal Microscopy List
> > [mailto:[hidden email]] On
> > Behalf Of Eric Scarfone
> > Sent: Friday, May 09, 2008 9:46 AM
> > To: [hidden email]
> > Subject: Re: interpolation Note of Caution
> >
> > Search the CONFOCAL archive at
> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
> >
> > Hello all
> > this problem certainly deserves a whole thread (it probably has
> > been
> > on this list before?).
> > I wonder how one should consider techniques such as background
> > subtraction that have been in use in video-microscopy even
> > before the
> > digital age!
> > Isn't it also manipulation?
> > Eric
> >
> > Eric Scarfone, PhD, CNRS,
> > Center for Hearing and communication Research
> > Department of Clinical Neuroscience
> > Karolinska Institutet
> >
> > Postal Address:
> > CFH, M1:02
> > Karolinska Hospital,
> > SE-171 76 Stockholm, Sweden
> >
> > Work: +46 (0)8-517 79343,
> > Cell: +46 (0)70 888 2352
> > Fax: +46 (0)8-301876
> >
> > email: [hidden email]
> > http://www.ki.se/cfh/
> >
> >
> > ----- Original Message -----
> > From: RICHARD BURRY <[hidden email]>
> > Date: Friday, May 9, 2008 3:28 pm
> > Subject: Re: interpolation Note of Caution
> > To: [hidden email]
> >
> > Search the CONFOCAL archive at
> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
> > > face="Times New
> > Roman">There are ethical limits as to what is
> > allowed in
> > manipulating micrographs. > yes">
> > Removing small unwanted objects is no different than
> > adding
> > small wanted objects.
> > A
> > great summary with examples is found in a Journal of Cell
> > Biology
> > article by Rossner and Yamada 2004 166:11-15 with the pdf
> > available at
> > http://www.jcb.org/cgi/reprint/166/1/11. > spacerun:
> > yes"> Journal editors are looking for these
> > modifications
> > because authors are misrepresentating their
> > data!<?xml:namespace
> > prefix = o ns = "urn:schemas-microsoft-
> > com:office:office" /><o:p></o:p>
> > Dick Burry Ohio State
> > University -----
> > Original Message ----- From: Zoltan Cseresnyes
> > <[hidden email]> Date: Friday, May 9, 2008 8:33
> > am Subject: Re: interpolation Note of Caution To:
> > [hidden email] > WEIGHT: normal;
> > FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">>
> > Search the CONFOCAL archive at
> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
> > I'm
> > completely with Jeremy on this subject, I don't think
> > it's good
> > practice to remove parts of your image just for its own
> > sake. You can e.g. false-colour the real or the artificial
> > pixels, in order to show the readers which objects to pay
> > attention
> > to. Just my 2c of course. > style="FONT-WEIGHT:
> > normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">> Zoltan 
>
> > <DIV class=gmail_quote> > FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > On
> > Fri, May 9, 2008 at 8:37 AM, Jeremy Adler <<A
> > href="java_script:main.compose
> ('new','t=[hidden email]')"
> > target="1">[hidden email]</A>> wrote: 
>
> > <BLOCKQUOTE class=gmail_quote style="PADDING-LEFT: 1ex;
> > MARGIN: 0px
> > 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid"> > WEIGHT:
> > normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">> Search the CONFOCAL archive
> > at <A
> > href="http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal"
> > target=1> > FONT-
> > STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?
> > S1=confocal</A> > normal; FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Removal of artefacts. > style="FONT-WEIGHT: normal;
> > FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">>
> > 1) Smooth your artefact free image, and use the mask
> > to select
> > the areas from the smoothed image that you wish to insert into
> > the
> > original. Only fiddle the problematic pixels. This will only
> > work for
> > small artefacts. 
> > > STYLE: normal;
> > BACKGROUND-COLOR: #f5f8f0">> 
> > However it
> > is worth asking why you wish wish to 'improve' the appearance of
> > your
> > images and whether this is ethical. > WEIGHT:
> > normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">> At the very least a full
> > description of why and how the published image differs from the
> > original image must be given. > normal;
> > FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">>
> > From your description you appear able to
> > decide
> > that some features are artefacts and define these features
> > sufficiently well to accurately generate a mask. If we
> > assume
> > that the artefects simply obliterate any underlying signal then
> > you
> > have no knowledge of what might have been found in those pixels.
> > And
> > no legitimate basis for 'improving' your image. <FO
> > NT style="FONT-WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE:
> > normal;
> > BACKGROUND-COLOR: #f5f8f0">> Zapping
> > the
> > artefacts and leaving clear and obvious blanks would be more
> > legitimate than 'improving' the original, but I would strongly
> > favour
> > publishing the originals and explaining/highlighting the
> > artefacts. > 14px; FONT-
> > STYLE: normal; BACKGROUND-COLOR: #f5f8f0">> Better
> > yet deal
> > with the source of the artefacts. > style="FONT-WEIGHT:
> > normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">> Jeremy Adler 
> > WEIGHT: normal;
> > FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">>
> > Cell Biology > normal; FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > The
> > Wenner-Gren Inst. > FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Arrhenius Laboratories E5 > style="FONT-WEIGHT: normal;
> > FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">>
> > Stockholm University > WEIGHT: normal; FONT-
> > SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Stockholm 106 91 > normal; FONT-
> > SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Sweden > normal; FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > ________________________________ > style="FONT-
> > WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-
> > COLOR:
> > #f5f8f0">> From: Confocal Microscopy List on behalf
> > of
> > MODEL, MICHAEL > SIZE: 14px;
> > FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Sent: Thu
> > 5/8/2008 16:09 > SIZE: 14px;
> > FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">> </
>
> > FONT>To: <A href="java_script:main.compose
> > ('new','t=[hidden email]')"
> > target="1">[hidden email]</A> > style="FONT-
> > WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-
> > COLOR:
> > #f5f8f0">> Subject:
> > inetrpolation > style="FONT-WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal;
> > BACKGROUND-COLOR: #f5f8f0">> Search the CONFOCAL
> > archive at
> > <A href="http://listserv.acsu.buffalo.edu/cgi-
> > bin/wa?S1=confocal"
> > target=1>http://listserv.acsu.buffalo.edu/cgi-bin/wa?
> > S1=confocal</A> > normal; FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Dear
> > List - > normal; FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Does
> > anyone know of a software (preferably an ImageJ plug-in) that
> > would
> > fill areas generated by a mask to make them merge smoothly with
> > the
> > surrounding areas? > T-WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal;
> > BACKGROUND-
> > COLOR: #f5f8f0">> For instance, a control image may
> > be taken
> > to identify artefacts. The question is, then, how to remove
> > small
> > unwanted objects from the main image without creating unsightly
> > holes.
> > Something like automatic "healing brush" in
> > Photoshop. > WEIGHT: normal; FONT-SIZE:
> > 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Michael Model, Ph.D. > WEIGHT: normal;
> > FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">>
> > Confocal Microscopy Core > style="FONT-WEIGHT:
> > normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">> Dpt. Biological
> > Sciences > WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-
> > COLOR:
> > #f5f8f0">> Kent State
> > University > WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COL
> > OR: #f5f8f0">> Kent, OH 44242 > style="FONT-
> > WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-
> > COLOR:
> > #f5f8f0">> tel. 330-672-
>
> > 2874 </BLOCKQUOTE> > clear=all> > style="FONT-WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal;
> > BACKGROUND-COLOR: #f5f8f0">> -- > style="FONT-
> > WEIGHT: normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-
> > COLOR:
> > #f5f8f0">> -- > normal; FONT-
> > SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Zoltan Cseresnyes > normal; FONT-
> > SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR: #f5f8f0">>
> > Facility manager, Imaging Suite > style="FONT-WEIGHT:
> > normal; FONT-SIZE: 14px; FONT-STYLE: normal; BACKGROUND-COLOR:
> > #f5f8f0">> Dept. of Zoology University of
> >
> >
> > --
> > BEGIN-ANTISPAM-VOTING-LINKS
> > ------------------------------------------------------
> >
> > Teach CanIt if this mail (ID 598043969) is spam:
> > Spam:
> > https://antispam.osu.edu/b.php?c=s&i=598043969&m=454cda2ec162Not
> > spam: https://antispam.osu.edu/b.php?
> c=n&i=598043969&m=454cda2ec162
> > Forget vote:
> > https://antispam.osu.edu/b.php?c=f&i=598043969&m=454cda2ec162----
> > --------------------------------------------------
> > END-ANTISPAM-VOTING-LINKS
>
> >
>
> Richard W. Burry, Ph.D.
> Department of Neuroscience, College of Medicine
> Campus Microscopy and Imaging Facility, Director
> The Ohio State University
> Associate Editor, Journal of Histochemistry and Cytochemistry
> 3018 Graves Hall
> 333 West Tenth Avenue
> Columbus, Ohio 43210
> Voice 614.292.2814 Cell 614.638.3345 Fax 614.688.8742