Protocol Online logo
Top : New Forum Archives (2009-): : Figure Making

Image beautification? - (Jun/07/2013 )

Is image beutification (changing image contrast/brightness/sharpness) the same as image fabrication (e.g. false information, adding or removing pixels, etc)?

Some papers now require authors to attach original images when they submit. So if this is the case, does it mean researchers have to try their best to capture the perfect image (from the microscope, chemiluminescence machine, camera) and not beautify it through post-processing? (Note: post-processing meaning perfecting images but not adding/removing pixels)

-science noob-

Well, I think almost everyone tries to capture the best image whether they want to photoshop it or not. If I am confident with my result but don't have a good image, I just write 'image not shown' in the article. If the reviewers accept then I'm lucky, but if they don't I'll try to find another way or repeat the experiment.

-Curtis-

A bit of enhancement seems to be fine - the problem starts when you start cutting out bands, filling stuff in with the paint brush etc.

-bob1-

since most of the image capture programs also offer "manipulation" in terms of reducing background, changing contrast, exposure time, camera sensitivity, etc. it should be no problem doing this also with other software...
Except the limits bob1 mentioned also cutting the image I'd see as problematic as you cannot see then where the wells where (running distance), if lanes with bands which you didn't want are cut away, or the size ladder could be replaced...so at least the original files have to be kept unchanged (perhaps with a hard copy from the printer).

-hobglobin-

I think you hit the nail on the head when you specified not adding or removing pixels as a good guideline for post-processing. As long as you are simply clarifying a result that already existed in the image, not creating a new result, you are generally in the clear. Also, processing should be as consistent as possible; for a series of histology pictures/blots/gels/whatever, one should endeavor to make identical changes to brightness/contrast for all images.

-gfischer-

hobglobin on Fri Jun 7 14:59:56 2013 said:


since most of the image capture programs also offer "manipulation" in terms of reducing background, changing contrast, exposure time, camera sensitivity, etc. it should be no problem doing this also with other software...
Except the limits bob1 mentioned also cutting the image I'd see as problematic as you cannot see then where the wells where (running distance), if lanes with bands which you didn't want are cut away, or the size ladder could be replaced...so at least the original files have to be kept unchanged (perhaps with a hard copy from the printer).


When I worked in the field of microscopic technology research, I noticed that all the papers had very nice images , no background noise etc..
The first days/weeks when I took pictures I got pretty upset and irritated because I was not able to get such nice images. I also used a top microscope, one of the most expensive and rare ones (only 5 in the world in use).
A few weeks later I received a file with some of my pictures (evaluated by a post doc) and all of a sudden they were nice...
They just "cleaned" them a bit... removed background etc...
I didnt like it to be honest, but I immediately know why those pics in the papers were so nice....

So where do you draw the line?

A stupid example: what if I run a PCR result on gel and my positive control is bad while my results and negative control are ok.
I cant use the picture of that gel..
So I run another gel: the + control is ok now, samples too, but negative control is bad..
What should I do? Cut and paste ? No, not legal... Send in the two pictures to state its ok..?
Or try again and run a third gel?

And what about papers that just mention it (in words) , that everything was ok.. and they dont show a picture (so no real prove).
And even if I have to send in a picture as proof, I can always "make up" a picture/proof by using fake samples/controls that are always ok...




bob1 on Fri Jun 7 09:06:05 2013 said:


A bit of enhancement seems to be fine - the problem starts when you start cutting out bands, filling stuff in with the paint brush etc.

Like I said above.... Where do you draw the line?


I really wonder how deep we should go in this..

To be honest, in my opinion good journals should start to work with professional reviewers that do nothing else than reviewing and are not related with an intitute anymore (this more objective) and more time to control it.
Some "reviewers" (professors) let their phd students controle the papers or post docs....

-pito-

pito on Fri Jun 7 19:17:51 2013 said:


bob1 on Fri Jun 7 09:06:05 2013 said:


A bit of enhancement seems to be fine - the problem starts when you start cutting out bands, filling stuff in with the paint brush etc.

Like I said above.... Where do you draw the line?


I really wonder how deep we should go in this..

To be honest, in my opinion good journals should start to work with professional reviewers that do nothing else than reviewing and are not related with an intitute anymore (this more objective) and more time to control it.
Some "reviewers" (professors) let their phd students controle the papers or post docs....

Well, basically it seems that so long as you declare the modifications that you have made, then there isn't a problem - I have seen figures in papers where they have clearly cut lanes from a membrane and pasted in others, but in the text it said something like "Image is a composite of several images" or something similar, and this has passed review. I find that, especially for microscope images, the brightness has to be greatly increased so that it will print well, but if you took an image that bright on the 'scope, you would have a completly saturated image with poor resolution of detail. Usually these sorts of things are constrained by the image itself - if you have to adjust the brightness and contrast too much the image starts to look obviously bad, so there is a sort of an internal limit to how much adjustment can be done.

-bob1-