Re: An alarming amount of image manipulation - time to fight back

Posted by Mario-2 on
URL: http://confocal-microscopy-list.275.s1.nabble.com/Re-Ana-alarming-amount-of-image-manipulation-tp592857p1313148.html

Re: An alarming amount of image manipulation - time to fig
(I apologize if people are getting this message twice. My problem is that my last couple of posts have not been acknowledged or sent back to me as part of the list. Thanks for your patience)


Oh Criminy,

I have tried to stay out of this. I now feel compelled to comment:

1. the human visual dynamic range under normal viewing conditions is much less that 8-bits (256 levels); its more like 6-bits (64 levels or worse). Thus, it is inevitable that microscopy images confocal or W.F. will require manipulation in intensity and contrast (and gamma correction) in order to generate prints or downloadable images that demonstrate the experimenters claims. It is rediculous that anyone should retract a paper because of minor intensity or contrast changes. It would be nice to know what procedures were carried out on an important data set and could be described in a few sentences. Further, I agree with advocates that original raw image data be made available by the journal and/or researchers upon request. In fact, uncompressed raw image files should be provided to the reviewers of a paper. Might save a lot time. As for Catherine Verfaillie and colleagues, without knowing the details, it seems like a black eye on common sense on the part of U of Minn. sorry don't mean to offend but consider item 2.

2.As for journals, different publishers use different settings for printing images. In my opinion, they must bear some of the responsibility for getting the images to print as the authors wish, but much of the time they tell the authors merely to enhance contrast because the journal prints tend to lose contrast and images appear washed out. I have experienced that even enhancing contrast to an absurd degree resulting in an image that is grossly inappropriate on my computer monitor, can still lead to inadequate contrast when printed in a journal! On the other hand, my experience with the journal Science was quite positive in that their Photoshop layout person worked with me to get the right print quality. I think it would be beneficial if all microscopy journals and others employing images did the same.

3. When it comes to creating fraudulent images free of manipulation artifacts, it is actually very time consuming and difficult, and I don't know why anyone would even bother. (Well actually, I do, but besides being immoral and unethical, it is not worth the trouble. You will be found out and your career will go down the drain). Anyway, before anybody gets paranoid, I know this from doing digital reconstructions of damaged historically important photographs (e.g., 90-100+ year old photogravures, metalochromes, platinum prints, etc). I try to create replicas of images that are as close to the originals as possible. Depending on the degree of damage or artifacts in an original print, it could take several weeks to be successful with a single 11"x17" original using the full image processing arsenal including Fourier analysis, building notch filters, etc.

Early in this thread it was mentioned that a histogram analysis can often reveal selective image tampering. I also think that this is a useful tool especially when it is applied with noise analysis on subsets of an image. Areas that have been manipulated by sharpening, blurring, selective intensity changes, and other types of local manipulation will create tell tale features that might not be visible to the eye but can be revealed by disproportionate statistical inconsistencies within the image as a whole. Systematic errors due to improperly adjusted PMT voltages or problems with amplifier gain dependent noise in a CCD camera can also be detected.


Mario

The issue of the recent request by the University of Minnesota to have Catherine Verfaillie retract her publication is worse than tragic: it has all the elements of a Puritanical witch-hunt, and all for the charge of changing the brightness level of three images.   Let me be quick to say that Morayma Reyes, the graduate student who was picked out for the "offense," changed brightness levels to conform the image for publication, exactly what a printer at the printing press would do (but somehow comes clean), and exactly what I advised her to do.
Believe me, no list of ethical rules by the Microscopy Society could EVER be written to have changed this outcome--which was clearly an investigation for political reasons--except one rule: no post-processing period (which is done anyway at the press: are they accountable, too?).

It is especially disheartening because Dr. Reyes is among the best researchers I have worked with.  I cannot forget the day when she called me into a room long after regular working hours and asked me to look through the microscope.  I saw a beating heart before this had ever been done with the use of stem cells.  I turned to her and commented, "This is too large for a rat heart and too small for a rabbit.  What is it?"  She smiled triumphantly and said, "We made it on a scaffold."

I agree wholeheartedly with Johan: fight back.  Especially for this so-called violation on research that has since been corroborated by other labs.
Jerry

-- 
________________________________________________________________________________
Mario M. Moronne, Ph.D.

[hidden email]
[hidden email]
[hidden email]