Source of Richardson test slide
Locked
12345
12345
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Dear All, This discussion could be endless. However, Science these days is rather disappointing than exciting. There are too many (boring) routine, daily results, but few exciting discoveries. In most cases people simply publish what they have. I am very sorry for "poor" students who have to dig this tons of rubbish. Or in other terms, there are plenty of Scientific Workers, but one has to be lucky to meet a Scientist. Thus, what to expect from a narrowly educated Scientific Worker? I haven't seen a lab where good physicists work together with biologists through a bridge of Math or cool Biophysics. Unfortunately, at this level we cannot change things. Cheers, Vitaly ----- Original Message ----- From: "Ingela Parmryd" <[hidden email]> To: <[hidden email]> Sent: Tuesday, June 24, 2008 7:42 AM Subject: SV: An alarming amount of image manipulation > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > Hi, > > The ease at which an operator can create colocalisation by changing the > brightness/contrast settings is one of the reasons why the overlay method > should not be used to assess colocalisation. Quantitation is a superior > alternative, for instance by the Pearson correlation coefficient. The PCC > is not sensitive to the actual intensities of the pixels but compares the > overall correlation between pixels from two channels and like for any > analysis, it is the raw images that should be analysed. > > Regards, > > Ingela Parmryd > > -----Ursprungligt meddelande----- > Från: Confocal Microscopy List [mailto:[hidden email]] För > Tina Carvalho > Skickat: den 24 juni 2008 00:07 > Till: [hidden email] > Ämne: Re: An alarming amount of image manipulation > > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > Hi, Gabor- > >> Do you think that your guidelines are sufficient? If someone is allowed >> to change the brightness/contrast and gamma settings for the different >> fluorescent channels separately and save those "new", edited images >> (especially if I do repeated cycles of this procedure) - it is pretty >> easy to create completely new images - showing e.g. more or less >> co-localization. The problem is that also the above mentioned procedures >> are able to manipulate the information content of an image... (OK they >> are always just reducing it - but still, one could use them to "cut out" >> the unwanted stuff). Don't you think so? > > Yes, I do think it's pretty easy to enhance or cut out the different > channels. (I usually find myself bringing up a channel rather than > reducing a channel.) But you are doing this with your choice of filters or > your voltage on your PMTs of any number of other ways already. So the > manipulation can begin before saving the image. Changing the levels on > each channel after the fact can be really useful to bring out information > that *is* really there but not captured and then mixed well because of > your type of camera or the health of your PMTs, the software is weird, or > that monitor is deficient, or whatever, so from that standpoint it is a > good thing. Or you can make up false colocalization, which is a big can of > worms anyway. There are all kinds of pitfalls along the way. In every case > I feel it is incumbent on the researcher to understand how the system is > working and what manipulations they are making, the consequences, and then > know what to report so that it can be checked or reproduced. > > Are the guidelines sufficient? No, but we couldn't think of a way to make > them any more easy to understand or implement. Unless you say that ANY > post-processing must be reported, including contrast and brightness and > levels. What do you think? > > Aloha, Tina > >> Greetings Gabor >> >> -- >> Gabor Csucs >> Light Microscopy Centre, ETH Zurich >> Schafmattstrasse 18, HPM F16 >> CH-8093, Zurich, Switzerland >> >> Web: www.lmc.ethz.ch >> Phone: +41 44 633 6221 >> Fax: +41 44 632 1298 >> e-mail: [hidden email] >> > > **************************************************************************** > * Tina (Weatherby) Carvalho * [hidden email] > * > * Biological Electron Microscope Facility * (808) 956-6251 > * > * University of Hawaii at Manoa * > http://www.pbrc.hawaii.edu/bemf* > **************************************************************************** > |
 
|
John Runions |
Re: An alarming amount of image manipulation
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Very constructive. I'm going to jump off a bridge now... Vitaly Boyko wrote: Search the CONFOCAL archive at --
*********************************
Visit
The Illuminated Plant
Cell
dot com |
|
|
Jerry Sedgewick-2 |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Peter Humphreys
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal I absolutely concur with Peter. To my way of thinking, 90% of the image quality deterioration occurs at the press. It seems that all but few journals work with agencies that know how to print largely black, and largely out-of-gamut primary colors. These require specific changes to colors and dynamic range in order to obtain adequate reproduction. Jerry Peter Humphreys wrote: > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > It would be quite helpful if journals could decide a clear and simple > standard for figures. Many papers are a little light on detail of the > manipulations used, but we frequently see vague and contradictory > information in the instructions for authors. > Since the journals usually insist on CMYK and we usually image RGB, the > colourspace conversion also alters the image. As it is almost impossible to > find which colour profile a journal uses, it is very difficult to get and > accurate conversion (photoshop usually does the conversion poorly); surely > the journals should take some of the responsibility for this part of the > publication process. Over the years I've seen the journals require more and > more formatting from our end. > > Peter > > > Peter Humphreys > Imaging Facility > Centre for Stem Cell Research > Cambridge > CB2 1QA > > > -----Original Message----- > From: Confocal Microscopy List [mailto:[hidden email]] On > Behalf Of Tina Carvalho > Sent: 24 June 2008 01:37 > To: [hidden email] > Subject: Re: An alarming amount of image manipulation > > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > >> As they say in Hungary (or so I was told), every good action gets its >> punishment. I think that people with a PhD, or about to get one, >> should know the difference between enhancing an image for display or >> publication, and misrepresentation (or fraud). If they are in doubt, >> they should also be smart enough to ask for advice. >> > > I agree here, as long as whomever they ask knows... > > By trying to > >> protect everybody against themselves, I am somewhat worried that out >> of all this, NIH or other agency will come up with a 1,500-page >> manual full with regulations and guidelines that will only make our >> lives just a little bit more complicated, probably with little impact >> on the amount of misconduct or cluelessness... >> > > Which is why MSA is trying (or at least I am, anyway) to come up with > guidelines that are easy to implement and make sense, to stave off further > complicated regulations. Input invited. > > >> I am currently analyzing cells labeled by FISH, and counting those >> that have nuclear and/or cytoplasmic staining. Nuclear spots are >> maybe 20-40 times brighter than the diffuse cytoplasmic signal. My >> eyes, camera, and computer monitor, all have different dynamic ranges >> and response curves. To see on the monitor what I see at the >> microscope, or to be able to print it, I need a pretty severe gamma >> adjustment to enhance the low intensities, otherwise I just will miss >> a lot of cells. This procedure however will change pixel intensities >> non-linearly, will not preserve intensity ratios between different >> regions of the image, and is pretty irreversible once applied. But >> that's OK... I know what I am doing and why I am doing it (and >> keeping the original data). On the other hand, without this gamma >> adjustment, the pictures I get on the screen (or paper) will just not >> match what I see at the microscope. We certainly don't want >> regulators telling us that non-linear contrast adjustment is no >> longer allowed. >> > > Same here, which is why I still think that our guidelines of being able to > adjust contrast, brightness, and levels/gamma makes sense. > > Recap: The Microscopy Society of America's whitepaper says you can adjust > brightness and contrast, and levels/gamma over the entire image. Anything > else should be reported as manipulation or enhancement. And you need to > store the "original" as uncompressed TIFF on archival media. We haven't > lately looked at other formats, like RAW, so the latter may change. Again, > if anyone wants to make an argument for another format, just dive in. > > Aloha, Tina > > >> -- >> Julio Vazquez, PhD >> Fred Hutchinson Cancer Research Center >> 1100 Fairview Ave. N., mailstop DE-512 >> Seattle, WA 98109-1024 >> >> >> http://www.fhcrc.org/ >> >> >> > > **************************************************************************** > * Tina (Weatherby) Carvalho * [hidden email] * > > * Biological Electron Microscope Facility * (808) 956-6251 * > * University of Hawaii at Manoa * http://www.pbrc.hawaii.edu/bemf* > > **************************************************************************** > > -- Jerry (Gerald) Sedgewick Program Director, Biomedical Image Processing Lab (BIPL) Department of Neuroscience, University of Minnesota 312 Church St. SE, 1-205 Hasselmo Hall Minneapolis, MN 55455 (612) 624-6607 [hidden email] http://www.bipl.umn.edu Author: "Scientific Imaging with Photoshop: Methods, Measurement and Output." Rawlight.com (dba Sedgewick Initiatives) 965 Cromwell Avenue Saint Paul, MN 55114 [hidden email] (651) 308-1466 http://www.quickphotoshop.com http://www.heartFROMstone.com http://www.rawlight.com --- Get FREE High Speed Internet from USFamily.Net! -- http://www.usfamily.net/mkt-freepromo.html --- |
|
|
Bill Oliver-3 |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Tina Carvalho
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal On Mon, 23 Jun 2008, Tina Carvalho wrote: > Back in the Really Old Days the development of negatives and the > production of photographic prints were detailed in the Materials and > Methods of a paper, right down to the concentration of developer, minutes, > etc., plus any dodging and burning. This eventually got shortened to just > negative and paper type and grade, and shortened more, until not much was > reported. Made for easier reading. I don't miss those days. BUT if we now > go back to more reporting of the details of image acquisition, it will > serve to promote reproducibility and, one hopes, make people think about > what they're doing so that they do not unintentially manipulate their > data. > The big problem, frankly, is that it is fundamentally impossible to create a digital image without image processing. The creation of the image itself requires exactly the kind of processing that some folk call fraud. The only difference is *where* you do the processing -- in the camera, in the computer, or in the display/printer. The implication is that image processing is OK as long as it's done ignorantly. The idea that there is some sort of pristine image in the camera sensor that must be preserved is analogous to claiming that it is fraud to develop and print photographic film because both the processing and the printing require "processing" the image -- dealing with issues of contrast, brightness, color balance, etc., etc., etc.. To claim that analogous processes in digital imagery is some sort of fraud is silly. Is color calibration "fraud?" I think not. In fact, I think a better argument can be made that there could be a greater misrepresentation without color calibration, gamma correction, etc. Attempts to say you can't do such things is merely requiring that the user be willfully ignorant of what's happening and use the default parameters (or last set parameters) of whatever imaging pipeline is used. As I've noted before, this was a hot topic in the forensic imaging world a few years ago. One set of best practices guidelines adopted by many forensics labs (the Scientific Working Group on Imaging Technologies, or SWGIT) guidelines are pretty explicit. See: http://www.theiai.org/guidelines/swgit/index.php and look at Section 11: Best Practices for Documenting Image Enhancement Basically the document divides enhancement methods into basic processes commonly used in image production and/or analogous to commonly accepted darkroom film processing (brightness/contrast adjustment, cropping, rotation, inversion, white balance, file format conversion, etc.) and "advanced" techniques such as deblurring, noise reduction, image restoration, etc. For basic processes, it is only necessary to indicate that they are done through some SOP or similar documentation. For advanced processing, all parts of the pipeline must be documented: "Documenting image enhancement steps should be sufficient to permit a comparably trained person to understand the steps taken, the techniques used, and to extract comparable information from the image. Documenting every change in every pixel value is discouraged because it adds nothing of value to the analysis. Exploratory enhancement operations not incorporated in the final image do not need to be documented. Test prints and/or intermediate images resulting from a variety of techniques not incorporated into the final image should be discarded. Minimum requirements for documentation include identifying the software application and/or techniques as well as the settings and parameters used. Automated processes, such as running user-defined macros, require only documenting usage if the process is defined in the agency documentation." billo |
|
|
Tina Carvalho |
Re: An alarming amount of image manipulation
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Right. So instead of saying you can't do anything, you just need to report everything, from pre-image acquisition to publication. That requirement is pretty straightforward. It may also ensure that investigators figure out how things work and what they're doing. Not a bad thing. Can we get everone on board with that? Aloha, Tina > The big problem, frankly, is that it is fundamentally impossible to create a digital image without image processing. The creation of the image itself requires exactly the kind of processing that some folk call fraud. The only difference is *where* you do the processing -- in the camera, in the computer, or in the display/printer. The implication is that image processing is OK as long as it's done ignorantly. > > The idea that there is some sort of pristine image in the camera sensor that must be preserved is analogous to claiming that it is fraud to develop and print photographic film because both the processing and the printing require "processing" the image -- dealing with issues of contrast, brightness, color balance, etc., etc., etc.. To claim that analogous processes in digital imagery is some sort of fraud is silly. Is color calibration "fraud?" I think not. In fact, I think a better argument can be made that there could be a greater misrepresentation without color calibration, gamma correction, etc. > > Attempts to say you can't do such things is merely requiring that the user be willfully ignorant of what's happening and use the default parameters (or last set parameters) of whatever imaging pipeline is used. > > As I've noted before, this was a hot topic in the forensic imaging world a few years ago. One set of best practices guidelines adopted by many forensics labs (the Scientific Working Group on Imaging Technologies, or SWGIT) guidelines are pretty explicit. See: > > http://www.theiai.org/guidelines/swgit/index.php > > and look at Section 11: Best Practices for Documenting Image Enhancement > > Basically the document divides enhancement methods into basic processes commonly used in image production and/or analogous to commonly accepted darkroom film processing (brightness/contrast adjustment, cropping, rotation, inversion, white balance, file format conversion, etc.) and "advanced" techniques such as deblurring, noise reduction, image restoration, etc. > > For basic processes, it is only necessary to indicate that they are done through some SOP or similar documentation. For advanced processing, all parts of the pipeline must be documented: > > "Documenting image enhancement steps should be sufficient to permit a comparably trained person to understand the steps taken, the techniques used, and to extract comparable information from the image. Documenting every change in every pixel value is discouraged because it adds nothing of value to the analysis. > > Exploratory enhancement operations not incorporated in the final image do not need to be documented. Test prints and/or intermediate images resulting from a variety of techniques not incorporated into the final image should be discarded. > > Minimum requirements for documentation include identifying the software application and/or techniques as well as the settings and parameters used. Automated processes, such as running user-defined macros, require only documenting usage if the process is > defined in the agency documentation." > > billo > **************************************************************************** * Tina (Weatherby) Carvalho * [hidden email] * * Biological Electron Microscope Facility * (808) 956-6251 * * University of Hawaii at Manoa * http://www.pbrc.hawaii.edu/bemf* **************************************************************************** |
|
|
Bill Oliver-3 |
Re: An alarming amount of image manipulation
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal On Tue, 24 Jun 2008, Tina Carvalho wrote: > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > Right. So instead of saying you can't do anything, you just need to report > everything, from pre-image acquisition to publication. That requirement is > pretty straightforward. It may also ensure that investigators figure out > how things work and what they're doing. Not a bad thing. Can we get > everone on board with that? > > Aloha, Tina > Well, not "everything." That's explicitly what we were tryintg to avoid -- in part to the demand by some attorneys that "everything" meant that you had to document every keystroke, every blind alley, every intermediate image, etc. There were a bunch of vendors pushing the idea that you had to use software that did logging so you could replay every little thing to counsel as part of discovery. We explicitly rejected that. Instead, we state that the documentation must only be to the detail so that a similarly-trained expert could come to the same conclusion. Frankly, I think it's silly to require that you document the chipset, the graphcis card, the brand and model number of your hard drive, etc. There has to be some reason involved, here. billo |
|
|
Rosemary.White |
change of thread-crossing borders/dealing with artefacts
|
|
In reply to this post by John Runions
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Re. the original thread, it’s not image manipulation that concerns me so much as not very good cell biology and people not recognising that they are publishing artefacts, either of staining, or GFP over-expression or mis-expression, etc. That’s the serious rubbish students are going to have to wade through. Sorry, I wasn’t going to add to this but couldn’t help myself.... At least John’s response raised a laugh at the start of the day.... cheers, Rosemary Dr Rosemary White [hidden email] CSIRO Plant Industry ph. 61 (0)2-6246 5475 GPO Box 1600 fax. 61 (0)2-6246 5334 Canberra, ACT 2601 Australia On 25/6/08 1:24 AM, "John Runions" <[hidden email]> wrote: Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal |
|
|
John Runions |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Bill Oliver-3
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi All, this has been an interesting and helpful discussion and I feel that it has highlighted many things for those of us that work in imaging to consider. Most of my colleagues however only work peripherally in imaging, i.e. they will use images or gels to illustrate their results but imaging is not their main activity. My take on all of this is that it is acceptable to do image-wide manipulation of levels, brightness/contrast, colour (as we did with printing from film)to get the 'best' image possible but that anything that feels illegitimate probably is illegitimate. When a colleague recently asked if I could remove a band from a gel image 'because it wasn't anything important and might just confuse the issue,' I felt ill (well, squeamish). I really believe that said colleague thought this was innocent and perfectly acceptable (I said no by the way). It is those colleagues and students that we need to educate and that education should be a part of the standard 'ethics' training that everyone should receive on an ongoing basis. John. Bill Oliver wrote: Search the CONFOCAL archive at --
*********************************
Visit
The Illuminated Plant
Cell
dot com |
|
|
James Beals |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Tina Carvalho
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Thanks Tina, This is the point, write it up in the methods. From the camera and the software that drives the camera, to the software and macros used to effect the images and measure the information. That way you are covered and other folks may learn techniques. I have seen too many methods stating, Images were enhanced for publication in ... (what ever software). This has been a very interesting thread. all the best James Beals [hidden email] 734.936-2051 205 Zina Pitcher Place 2038 MBNI Molecular and Behavioral Neuroscience Institute University of Michigan Ann Arbor, Mi 48109 On Jun 24, 2008, at 5:13 PM, Tina Carvalho wrote: > Right. So instead of saying you can't do anything, you just need to > report > everything, from pre-image acquisition to publication. That > requirement is > pretty straightforward. It may also ensure that investigators > figure out > how things work and what they're doing. Not a bad thing. Can we get > everone on board with that? > > Aloha, Tina |
|
|
Daniel James White |
Re: An alarming amount of image manipulation
|
|
In reply to this post by John Oreopoulos
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Vitaly, You should come see an example of "a lab where good physicists work together with biologists through a > > bridge of Math or cool Biophysics" here at MPI-CBG in Dresden. That's exactly what we do have and are pushing to have more of. We even have a new software engineering facility, in a Cell/ Developmental Biology lab. ...similar things happen at Janelia Farms. Dan On Jun 25, 2008, at 6:00 AM, CONFOCAL automatic digest system wrote: > Date: Tue, 24 Jun 2008 10:52:23 -0500 > From: Vitaly Boyko <[hidden email]> > Subject: Re: An alarming amount of image manipulation > > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > Dear All, > > This discussion could be endless. > > However, Science these days is rather disappointing than exciting. > > There are too many (boring) routine, daily results, but few exciting > discoveries. In most cases people simply publish what they have. I > am very > sorry for "poor" students who have to dig this tons of rubbish. > Or in other terms, there are plenty of Scientific Workers, but one > has to be > lucky to meet a Scientist. > > Thus, what to expect from a narrowly educated Scientific Worker? I > haven't > seen a lab where good physicists work together with biologists > through a > bridge of Math or cool Biophysics. > > Unfortunately, at this level we cannot change things. > > Cheers, > > Vitaly Dr. Daniel James White BSc. (Hons.) PhD Senior Microscopist / Image Processing and Analysis Light Microscopy Facility Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany New Mobile Number!!! +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) http://www.bioimagexd.net http://www.chalkie.org.uk [hidden email] ( [hidden email] ) |
|
|
Michael Cammer |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Bill Oliver-3
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal The big problem, frankly, is that it is fundamentally impossible to create a any image without image processing. The creation of any image itself requires exactly the kind of processing that some folk call fraud. Every image is a human construct. Discussions such as this get bogged down in the details of whether a bicubic or bilinear interpolation is more true when the real questions are more fundamental, such as do the pictures accurately illustrate for colleagues or are the images showing a new biological (or pick your field phenomenon) process and are not the result of the instrument itself. When the first telescopes and microscopes were invented, some people didn't believe that what they saw was a reality out in the sky or in a drop of fluid; they said that the new instruments created the images. Lesson: understand the technologies you use to study and discover. Also, the fact is that the scientist who tells the story in the sexiest way wins. Is the story itself one of merit? That's of far more import than whether the images are silver grains, pixels or hand drawn sketches. We all prefer aesthetically pleasing images; do they represent the real in a justifiable manner? Yes, we do need norms or guidelines for honesty and adherence to a system or paradigm for assessing "truth", but this nitpicky stuff is just way too, well, nitpicky. -Michael >The big problem, frankly, is that it is fundamentally impossible to >create a digital image without image processing. The creation of >the image itself requires exactly the kind of processing that some >folk call fraud. The only difference is *where* you do the >processing -- in the camera, in the computer, or in the >display/printer. The implication is that image processing is OK as >long as it's done ignorantly. > >The idea that there is some sort of pristine image in the camera >sensor that must be preserved is analogous to claiming that it is >fraud to develop and print photographic film because both the >processing and the printing require "processing" the image -- >dealing with issues of contrast, brightness, color balance, etc., >etc., etc.. To claim that analogous processes in digital imagery is >some sort of fraud is silly. Is color calibration "fraud?" I think >not. In fact, I think a better argument can be made that there >could be a greater misrepresentation without color calibration, >gamma correction, etc. > >Attempts to say you can't do such things is merely requiring that >the user be willfully ignorant of what's happening and use the >default parameters (or last set parameters) of whatever imaging >pipeline is used. > >As I've noted before, this was a hot topic in the forensic imaging >world a few years ago. One set of best practices guidelines adopted >by many forensics labs (the Scientific Working Group on Imaging >Technologies, or SWGIT) guidelines are pretty explicit. See: > >http://www.theiai.org/guidelines/swgit/index.php > >and look at Section 11: Best Practices for Documenting Image Enhancement > >Basically the document divides enhancement methods into basic >processes commonly used in image production and/or analogous to >commonly accepted darkroom film processing (brightness/contrast >adjustment, cropping, rotation, inversion, white balance, file >format conversion, etc.) and "advanced" techniques such as >deblurring, noise reduction, image restoration, etc. > >For basic processes, it is only necessary to indicate that they are >done through some SOP or similar documentation. For advanced >processing, all parts of the pipeline must be documented: > >"Documenting image enhancement steps should be sufficient to permit >a comparably trained person to understand the steps taken, the >techniques used, and to extract comparable information from the >image. Documenting every change in every pixel value is discouraged >because it adds nothing of value to the analysis. > >Exploratory enhancement operations not incorporated in the final >image do not need to be documented. Test prints and/or intermediate >images resulting from a variety of techniques not incorporated into >the final image should be discarded. > >Minimum requirements for documentation include identifying the >software application and/or techniques as well as the settings and >parameters used. Automated processes, such as running user-defined >macros, require only documenting usage if the process is >defined in the agency documentation." > >billo ____________________________________________________________________________ Michael Cammer, Senior Light Microscopist, Analytical Imaging Facility, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461 URLs: microscopy http://www.aecom.yu.edu/aif/ and art http://coxcammer.com/ |
|
|
Michael Herron |
Re: An alarming amount of image manipulation
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
I concur. Guidelines might help, but committees, and regulations carry a real cost in time, dollars, and the roads not taken. In the end, experiments must be confirmed independently by replication in other laboratories. We will never regulate away mistakes or subterfuge.
Mike
On Jun 25, 2008, at 11:45 AM, Michael Cammer wrote:
Michael J. Herron, U of MN, Dept. of Entomology 612-624-3688 (office) 612-625-5299 (FAX) |
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Aloha Tina and others, I’m interested in hearing more about how I can help MSA with
this issue. I have had a set of digital imaging guidelines online since
2001. The guidelines were originally written for the students and staff that
I support, but I think they are more broadly applicable. I’ve presented
them at a couple of meetings in the last few years. The guidelines can be
found at this web address in either HTML or PDF format: http://swehsc.pharmacy.arizona.edu/exppath/micro/digimage_ethics.html I have a review article on this topic sitting in the queue at
the Journal of Science and Engineering Ethics. Unfortunately it doesn’t look
like it will see the light of day until sometime mid-next year. Also, I collaborated with the good folks at the Center for
Ethics and Values in the Sciences (Univ of Alabama – Birmingham) on an Office
of Research Integrity grant to develop a web site on this topic.
The site uses my 12 guidelines and a video-based case study of a student who
discovers that her great confocal colocalization images were really spectral
bleed-through. Then she learns that the post-doc she reports to is not at
all interested in doing anything about the fact that this “technicality”
invalidates an image in a recently submitted paper from the fictitious
lab. The web site is still being finished, but I’ve been told that it
will be completed this summer. The URL is: http://www.uab.edu/researchintegrityandimages Sorry I haven’t chimed in earlier, I’m finishing up a stay at
Jim Pawley’s 3D Microscopy of living cells course and I’ve been rather busy. Yours, Doug Cromey ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 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] voice: 520-626-2824 fax: 520-626-2097 http://swehsc.pharmacy.arizona.edu/exppath/ Home of: "Microscopy and Imaging Resources on the WWW" |
|
|
Claire Brown |
Re: An alarming amount of image manipulation
|
|
In reply to this post by John Oreopoulos
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal I wanted to add an interesting situation to the list of image manipulation issues. I was making figures for my imaging pitfalls paper for J. Cell Sci. last year and I was trying to show undersampling. So I used a CCD camera and did 1x1, 2x2, 3x3, and 4x4 binning of the camera. Of course at 60X with 4x4 binning you can clearly see the squares on the images. I could not open the 4x4 binned image in Photoshop because it kept smoothing the image to make it the dpi of the default images. So what I did is I put it in MetaMorph, made the image large and did a duplicate as displayed and saved it as a jpg. Then when I took it into photoshop it looked "correct". A few months later I got the poster proofs for the article and to my surprise the 1x1, 2x2, 3x3 and 4x4 binned images all looked identical and no longer made my point of undersampling because they were all smoothed out. Of course they probably print the poster at 600 dpi or higher so I assume this caused the smoothing. I ended up on the phone for about an hour with the printing production manager and what we ended up having to do is select those regions of the poster and set the printing dpi very low (50 dpi or lower for that part of the poster). So the images on the poster were smoothed and then pixelated so they are not quite correct but they still make my point. I have a copy of a power point presentation for the pitfalls poster and it includes the images before and from the proofs if anyone is interested in seeing them. Go to http://www.lifesciencescomplex.mcgill.ca/imaging/links/Workshops2007 and download Microscopy-pitfalls-talk.ppt. Sometimes it is very difficult NOT to manipulate your data! Sincerely, Claire |
|
|
RICHARD BURRY |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Tina Carvalho
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Slippery slope time. The integrity of the scientist presenting a micrograph should allow us to trust the micrographs as accurately representing the data. While there are guidelines for images, the results presented in papers are based on the author’s conclusions and falsification is difficult to detect. In the end, trust is what keeps scientific communication at such a reliable level. Rather than take the police approach to find altered images, a more positive approach would be education. This would mean setting guidelines and explaining the reasons for them. I have worked with scientists that did not intend to misrepresent their data, but did not know enough about microscopy. To have all the micrographs in a figure show the same level of background, some software adjustments may work, but in many cases if the micrographs are different enough, it will be necessary to go back to the microscope and collect additional images. The bottom line is that we must honestly present our data and not mislead our colleagues. Dick ----- Original Message ----- From: Tina Carvalho <[hidden email]> Date: Monday, June 23, 2008 6:07 pm Subject: Re: An alarming amount of image manipulation To: [hidden email] > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > Hi, Gabor- > > > Do you think that your guidelines are sufficient? If someone > is allowed > > to change the brightness/contrast and gamma settings for the > different > > fluorescent channels separately and save those "new", edited > images > > (especially if I do repeated cycles of this procedure) - it is > pretty > > easy to create completely new images - showing e.g. more or > less > > co-localization. The problem is that also the above mentioned > procedures > > are able to manipulate the information content of an image... > (OK they > > are always just reducing it - but still, one could use them to > "cut out" > > the unwanted stuff). Don't you think so? > > Yes, I do think it's pretty easy to enhance or cut out the different > channels. (I usually find myself bringing up a channel rather than > reducing a channel.) But you are doing this with your choice of > filters or > your voltage on your PMTs of any number of other ways already. > So the > manipulation can begin before saving the image. Changing the > levels on > each channel after the fact can be really useful to bring out > informationthat *is* really there but not captured and then > mixed well because of > your type of camera or the health of your PMTs, the software is > weird, or > that monitor is deficient, or whatever, so from that standpoint > it is a > good thing. Or you can make up false colocalization, which is a > big can of > worms anyway. There are all kinds of pitfalls along the way. In > every case > I feel it is incumbent on the researcher to understand how the > system is > working and what manipulations they are making, the > consequences, and then > know what to report so that it can be checked or reproduced. > > Are the guidelines sufficient? No, but we couldn't think of a > way to make > them any more easy to understand or implement. Unless you say > that ANY > post-processing must be reported, including contrast and > brightness and > levels. What do you think? > > Aloha, Tina > > > Greetings Gabor > > > > -- > > Gabor Csucs > > Light Microscopy Centre, ETH Zurich > > Schafmattstrasse 18, HPM F16 > > CH-8093, Zurich, Switzerland > > > > Web: www.lmc.ethz.ch > > Phone: +41 44 633 6221 > > Fax: +41 44 632 1298 > > e-mail: [hidden email] > > > > **************************************************************************** > * Tina (Weatherby) > Carvalho * [hidden email] * > * Biological Electron Microscope Facility * (808) 956- > 6251 * > * University of Hawaii at > Manoa * http://www.pbrc.hawaii.edu/bemf* > **************************************************************************** > > > -- > BEGIN-ANTISPAM-VOTING-LINKS > ------------------------------------------------------ > > Teach CanIt if this mail (ID 615327917) is spam: > Spam: > https://antispam.osu.edu/b.php?c=s&i=615327917&m=f2a12710f27bNot > spam: https://antispam.osu.edu/b.php?c=n&i=615327917&m=f2a12710f27b > Forget vote: > https://antispam.osu.edu/b.php?c=f&i=615327917&m=f2a12710f27b---- > -------------------------------------------------- > 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 277 Biomedical Research Tower 460 West Twelfth Avenue Columbus, Ohio 43210 Voice 614.292.2814 Cell 614.638.3345 Fax 614.247.8849 |
|
|
RICHARD BURRY |
Re: An alarming amount of image manipulation
|
|
In reply to this post by Tina Carvalho
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Slippery slope time. The integrity of the scientist presenting a micrograph should allow us to trust the micrographs as accurately representing the data. While there are guidelines for images, the results presented in papers are based on the author’s conclusions and falsification is difficult to detect. In the end, trust is what keeps scientific communication at such a reliable level. Rather than take the police approach to find altered images, a more positive approach would be education. This would mean setting guidelines and explaining the reasons for them. I have worked with scientists that did not intend to misrepresent their data, but did not know enough about microscopy. To have all the micrographs in a figure show the same level of background, some software adjustments may work, but in many cases if the micrographs are different enough, it will be necessary to go back to the microscope and collect additional images. The bottom line is that we must honestly present our data and not mislead our colleagues. Dick ----- Original Message ----- From: Tina Carvalho <[hidden email]> Date: Monday, June 23, 2008 6:17 pm Subject: Re: An alarming amount of image manipulation To: [hidden email] > Search the CONFOCAL archive at > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > Hi, Steve- > > One reason I got into this is that I can make an entirely new > species of > insect or copepod in Photoshop and probably you wouldn't be able > to detect > it...! > > It does boil down to the integrity of the researcher, but > increasingly the > researcher has no idea what their actions during image > acquisition and > post-processing for publication mean for the integrity of the > data. Most > of the problems I see here are from people being uninformed, not > malicious. So I agree education and training are the solution. > I'm doing > my personal best, but I can't vouch for anyone else on this > campus! We've > got people who did not grow up with digital images mentoring > students who > have never known anything different and are used to just > clicking on > filters to get a pretty picture. If they come through my > facility, they > get to hear me rant about it and, I hope, go away a little more > cautious. How to do this on a large scale? More explicit > guidelines for > submissions to journals? Anyone want to help out with this? Volunteers > welcome. Join the Committee on the Ethics of Digital Imaging in the > Microscopy Society of America. > > Aloha, Tina > > > I guess I'll add my two cents here. As I've said in the past > (and > > this was true back in the darkroom days) it still comes down > to the > > integrity of the investigator preparing the images. There are > just > > about always ways to get around any safeguards against image > fraud > > that are put in place. For example, if someone were to do > an > > excellent job of producing a fraudulent image in photoshop, > print it > > out, and then retake a digital picture of the figure, how easy > would > > it be to detect manipulation? I realize that in certain cases > this > > might be possible, but in other cases it would probably be > very > > difficult. > > > > I tend to think that good education and training of > investigators, > > and subsequently, their students is the best route to go. > > > > Just some thoughts, > > > > Steve > > > > On Jun 23, 2008, at 4:17 PM, Csucs Gabor wrote: > > > > > Search the CONFOCAL archive at > > > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal > > > > > > Dear Tina, > > > > > > Do you think that your guidelines are sufficient? If someone > is > > > allowed to change the brightness/contrast and gamma settings > for > > > the different fluorescent channels separately and save those > "new", > > > edited images (especially if I do repeated cycles of > this > > > procedure) - it is pretty easy to create completely new > images - > > > showing e.g. more or less co-localization. The problem is > that also > > > the above mentioned procedures are able to manipulate > the > > > information content of an image... (OK they are always > just > > > reducing it - but still, one could use them to "cut out" > the > > > unwanted stuff). Don't you think so? > > > > > > Greetings Gabor > > > > > > -- > > > Gabor Csucs Light Microscopy Centre, ETH Zurich > > > Schafmattstrasse 18, HPM F16 CH-8093, Zurich, Switzerland > > > > > > Web: www.lmc.ethz.ch > > > Phone: +41 44 633 6221 > > > Fax: +41 44 632 1298 > > > e-mail: [hidden email] > > > > **************************************************************************** > * Tina (Weatherby) > Carvalho * [hidden email] * > * Biological Electron Microscope Facility * (808) 956- > 6251 * > * University of Hawaii at > Manoa * http://www.pbrc.hawaii.edu/bemf* > **************************************************************************** > > > -- > BEGIN-ANTISPAM-VOTING-LINKS > ------------------------------------------------------ > > Teach CanIt if this mail (ID 615333472) is spam: > Spam: > https://antispam.osu.edu/b.php?c=s&i=615333472&m=18999c8684d8Not > spam: https://antispam.osu.edu/b.php?c=n&i=615333472&m=18999c8684d8 > Forget vote: > https://antispam.osu.edu/b.php?c=f&i=615333472&m=18999c8684d8---- > -------------------------------------------------- > 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 277 Biomedical Research Tower 460 West Twelfth Avenue Columbus, Ohio 43210 Voice 614.292.2814 Cell 614.638.3345 Fax 614.247.8849 |
|
|
George McNamara |
Re: An alarming amount of image manipulation
|
|
In reply to this post by cromey
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi Doug,
No amount of MSA or JCB or ORI guidelines will handle someone who uses the phrase, "data not shown" for results they may or may not have generated. My advice for you and the MSA and everyone else is to for all the raw data that went into the study to be posted online at the journal web site as part of the publication. The OME crowd completely missed the boat about minimal information/metadata in their recent MISFISHIE proposal (PubMed 18327244). Any of the MISFISHIE authors reading this: the first sentence of this paragraph is my feedback. The cost of a 1 Terabyte drive is currently about $250 (and would be a lot less for the publishers buying in bulk). The typical output of an NIH funded lab is 1 to 2 peer reviewed primary research publications per year per $250,000 RO1 grant (my guess, far less than 10 Gigabytes of data per primary publication, or about $2.50 in space). Plus, this way when the lab's or your confocal microscope computer's hard drive dies, there will be a full record of the data online. Enjoy, George At 09:12 AM 6/26/2008, you wrote: Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal |
|
|
Jason Swedlow |
Re: An alarming amount of image manipulation
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi All-
George's comments are very important-- I'm happy to respond for OME. The MIFISHIE group is a large group and perhaps someone else should speak for it. George, your idea is potentially very useful-- publish all raw data with the paper. You are right-- cost of storage really isn't the limiting problem (although you ignore the cost of duplication and backup, which is essential for any real distribution of published data). The proliferation of proprietary file formats makes access to these files difficult at best. Moreover, you ignore the very real problem that simply downloading that data is potentially a huge burden that doesn't scale well (any good paper will have 20 different people in my own institute downloading the same data-- so now we have 20 TB filling various storage systems). Most importantly, that raw data doesn't capture the experimental protocols, analysis, annotations, ROIs, measurements, etc. that give that data meaning. These are the problems we are trying to address in the OME project. There is no question we are taking steps in this direction, and have not yet solved the full problem. I will argue we are getting there. Am happy to discuss that, but most probably should be on a different thread. Alternatively, visit our web site (http://openmicroscopy.org) where we keep alot of this info. In general, though, George is correct-- if "real data" is published, it is much harder-- but by no means impossible-- to inappropriately manipulate data. Full access to data in support of publication is a goal we strongly support. Cheers, Jason On Sun, Jun 29, 2008 at 10:29 PM, George McNamara <[hidden email]> wrote:
-- ************************** Wellcome Trust Centre for Gene Regulation & Expression College of Life Sciences MSI/WTB/JBC Complex University of Dundee Dow Street Dundee DD1 5EH United Kingdom phone (01382) 385819 Intl phone: 44 1382 385819 FAX (01382) 388072 email: [hidden email] Lab Page: http://www.dundee.ac.uk/lifesciences/swedlow/ Open Microscopy Environment: http://openmicroscopy.org ************************** |
|
|
Robert Peterson-3-3 |
Re: An alarming amount of image manipulation
|
|
I haven't followed all of this thread, so excuse me if someone has already mentioned this. One problem not addressed here is the manipulation of data capture itself. I have trained a number of people with previous confocal experience who have no idea what I am talking about when I pull up the range indicator to show saturation levels. When I allow them to do the set-up themselves it is often over-saturated and the offset is overdone as well.
Posting this "original" data won't help since it is already tweaked, unless it allows a reviewer to open the image meta data and see how it was collected and ask for better images. Just my two-cents. Robert Sent from my Verizon Wireless BlackBerry From: Jason Swedlow <[hidden email]> Date: Mon, 30 Jun 2008 13:42:20 +0100 To: <[hidden email]> Subject: Re: An alarming amount of image manipulation Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Hi All- George's comments are very important-- I'm happy to respond for OME. The MIFISHIE group is a large group and perhaps someone else should speak for it. George, your idea is potentially very useful-- publish all raw data with the paper. You are right-- cost of storage really isn't the limiting problem (although you ignore the cost of duplication and backup, which is essential for any real distribution of published data). The proliferation of proprietary file formats makes access to these files difficult at best. Moreover, you ignore the very real problem that simply downloading that data is potentially a huge burden that doesn't scale well (any good paper will have 20 different people in my own institute downloading the same data-- so now we have 20 TB filling various storage systems). Most importantly, that raw data doesn't capture the experimental protocols, analysis, annotations, ROIs, measurements, etc. that give that data meaning. These are the problems we are trying to address in the OME project. There is no question we are taking steps in this direction, and have not yet solved the full problem. I will argue we are getting there. Am happy to discuss that, but most probably should be on a different thread. Alternatively, visit our web site (http://openmicroscopy.org) where we keep alot of this info. In general, though, George is correct-- if "real data" is published, it is much harder-- but by no means impossible-- to inappropriately manipulate data. Full access to data in support of publication is a goal we strongly support. Cheers, Jason On Sun, Jun 29, 2008 at 10:29 PM, George McNamara <[hidden email]> wrote:
-- ************************** Wellcome Trust Centre for Gene Regulation & Expression College of Life Sciences MSI/WTB/JBC Complex University of Dundee Dow Street Dundee DD1 5EH United Kingdom phone (01382) 385819 Intl phone: 44 1382 385819 FAX (01382) 388072 email: [hidden email] Lab Page: http://www.dundee.ac.uk/lifesciences/swedlow/ Open Microscopy Environment: http://openmicroscopy.org ************************** |
|
|
Mayandi Sivaguru |
Re: An alarming amount of image manipulation
|
|
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
I think it is a great idea to post or let the raw data accessible to the readers, this can be done from the institution side too instead of the journal. In addition, to address the image acquisition, analysis and presentation methods the authors used, the journal could provide free access to programs like this "Easy Screen Capture" http://www.easyscreencapturevideo.com/, this can be installed in any computer and the user/author can create a video together with audio (if necessary), how the image was acquired (directly from the LSM or Microscope software while the sample is on the stage), and how the analysis are performed and how the images are finally put together in the presentation software, all could be captured step by step and then could be uploaded to the journal website together as the web only data. This is not only helpful for the reviewers and readers as well as helpful in teaching students of the particular method or technique involved/demonstrated in the paper. Finally, it is great for the microscopy facilities to post different methods and techniques used to operate complex systems, software and techniques. Shiv At 08:07 AM 6/30/2008, Robert E. Peterson wrote: I haven't followed all of this thread, so excuse me if someone has already mentioned this. One problem not addressed here is the manipulation of data capture itself. I have trained a number of people with previous confocal experience who have no idea what I am talking about when I pull up the range indicator to show saturation levels. When I allow them to do the set-up themselves it is often over-saturated and the offset is overdone as well. Posting this "original" data won't help since it is already tweaked, unless it allows a reviewer to open the image meta data and see how it was collected and ask for better images. Just my two-cents. Robert Microscopy Facility Manager 8, Institute for Genomic Biology University of Illinois at Urbana-Champaign 1206 West Gregory Dr. Urbana, IL 61801 USA Office: 217.333.1214 Fax: 217.244.2496 [hidden email] http://core.igb.uiuc.edu |
«
Return to Confocal Microscopy List
|
1 view|%1 views
| Free forum by Nabble | Edit this page |