Found my first image duplication in peer review.
Exciting! Depressing!

What next, do I even bother completing the review or do I stop here and flag it to the editor?

#ImageForensics #plantscience #academicchatter

Image sleuth @ElisabethBik is back on Mastodon 🕵️‍♂️ :mastodon: 🎉

https://med-mastodon.com/@ElisabethBik/110969401224111581

#ImageForensics #ImageManipulation #PublicationEthics #ResearchIntegrity #AcademicMastodon

#ImageForensics comes to #Mastodon 🙂
How many duplications can you find?

The recent wave of pope-related AI images, and the accompanying hot takes about whether or not we've now finally left an era of »visual truth« made me think about the relationship between two modes of online image interpretation: #WildForensis and #InstantMemeification.

Popular versions of #imageforensics have been a staple of social media for some time: People just love to speculate about whether or not a widely shared image has been manipulated. AI images in their current form are a perfect object of such »wild forensis«: Clues that an image was generated are now often so subtle that they're only visible at second glance, but you still don't need any advanced technical skills to find them.

At the same time, AI #imagegeneration seems to be the perfect meme machine: The way it transforms written concepts into memorable images already largely corresponds to the combinatorial logic of memes. What’s more, AI makes it extremely easy to produce endless variations of an already mass-circulated image, effectively combining recognizability and unpredictability.

As social media phenomena, both #WildForensis and #InstantMemeification are part of an online reaction economy, ways of responding to widely circulated images, albeit in almost opposite ways: Either by focusing on meaningless details or by rearranging semantic content. Nevertheless, in many cases both seem related, almost intertwined and reinforcing, resulting in what might be called »image reaction chains«.

So if you were to ask me what these pope-related AI images tell us about our cultural moment, it has nothing to do with »visual truth,« which has always been little more than a myth. Rather, we now see how #ImageReactionChains work in social media, how they are driven both by both the collective desire to discover forensic clues and the joy of reinterpreting semantic content, and how they are now transformed through AI image generation.

#ImageForensics challenge.
ImageTwin found three overlapping sets of panels. Find at least two and be one of the first to qualify for the Emoji Awards!

#ImageForensics
Seven panels, each from a different treatment group. All panels should be unique, without any overlaps. But, can you find some unexpected overlaps?

They could have gotten a collection of green spots from *anywhere*, but they chose to copy a panel from the same figure. I wonder what level of fraud out there goes completely unnoticed, just because authors are better cheats.
---
RT @MicrobiomDigest
Who can spot an unexpected overlap in these panels?
#ImageForensics https://t.co/ljaQ20qSG6
https://twitter.com/MicrobiomDigest/status/1619957089610366977

#ImageForensics
Twelve panels representing 4 treatment groups and 3 different proteins.
Each panel should be unique and not overlap with other panels.
But can you spot an overlap?

#ImageForensics challenge.

Two Western blots (each showing a different protein), performed on samples from five different treatment groups (rats).

Can you spot some unexpected similar-looking bands?

#ImageForensics challenge.
Eight panels, each showing a different experiment. Migration vs invasion, two different cell lines, and two different treatments.
However, some panels show unexpected overlaps. Which ones?

#ImageForensics challenge!

Nine panels of rat brain tissue of three treatment groups, stained with antibodies against 3 different proteins.

None of these panels are expected to overlap. Yet, two panels shown an overlap.

Which two?

#ImageForensics.

In this figure from a peer-reviewed paper (cited 26 times), you see 10 panels representing a control and 4 time points, and two different immunostainings, each staining for a different protein.

But. Three pairs of panels overlap with each other.

Challenge: Find the overlapping pairs of panels.

If you are the first to find at least two pairs, you can win an Emoji Award.

Can you find some serious problems here?
These are cryoanalytical electron microscope images of hippocampal neurons, taken by an NIH lab.
#ImageForensics

#ImageForensics
Brain tissue of mice from different treatment groups and timepoints. Can you spot an unexpected overlap?

An easy #ImageForensics for you.
Published in a journal once known for strict screening of images before papers got published. Now it's up to the unpaid image sleuths to detect these.
Can you detect the problem in these Western blots?

Photos of bacteria treated with leaf extract to see if the leafs can inhibit their growth.
Can you spot a duplication or overlap?
#ImageForensics #Pseudomonas

@chartgerink the moment you said this, it made me think of @ElisabethBik's #ImageForensics. Elisabeth, please share any pointers if you'd have the time (or join us! 😉 😄 )

Now that I've moved to @tldr.nettime, maybe I should rehash my #introduction:
Hi everyone, my name is Roland, I am Berlin-based media and visual culture scholar interested in the history, theory, and aesthetics of #operativeimages. Among other stuff, I have written on the (pre-)history of #facialrecognition & its contemporary implications.
Currently, I am a researcher at the Bochum SFB »Virtual Lifeworlds«, working in a project on virtual #imagearchives. Besides that, am also interested in #imagegeneration, #imageforensics, and #urbansignage.

Stanford president’s research under investigation for scientific misconduct, University admits ‘mistakes’
#ImageForensics
* Paper by University President Marc Tessier-Lavigne is now under investigation
* Theo Baker writes at The Stanford Daily.
Bik, though, disagreed with the university’s assertion that the “mistakes” do not impact the scientific integrity of the papers. “I do not agree with [the] statement that these issues have no bearing on the data or the results"
https://stanforddaily.com/2022/11/29/stanford-presidents-research-under-investigation-for-scientific-misconduct-university-admits-mistakes/

Can you spot the two overlapping panels?
#ImageForensics