Latest posts tagged with #ImageForensics on Bluesky
A collage of four mice, lying on their backs on a raster. Each mouse has a blue signal on their belly, with different shades of blue and cyan indicating different luminescence signal strengths.
Can you spot the #ImageForensics problem here?
(and yes, the taped mice are problematic, too).
Figure 2, screenshot. We see 2 rows with 3 panels each. Panels show grey background and white little dots, bacterial colonies. I have marked the top left and top middle panels with blue boxes, because they are identical. Legend reads: "Figure 2. The D-alanine demand identification of recombinant L. casei strain pPG-Alr-COE/∆Alr Figure 2. The D-alanine demand identification of recombinant L. casei strain pPG-Alr-COE/ΔAlr W56. The recombinant strains pPG-Alr-COE/ΔAlr W56, pPG-T7g10-PPT/ΔAlr W56, and ΔAlr W56 W56. The recombinant strains pPG-Alr-COE/∆Alr W56, pPG-T7g10-PPT/∆Alr W56, and ∆Alr W56 were streaked on de Man–Rogosa–Sharpe (MRS) plates to detect the demand of D-alanine. “+” were streaked on de Man-Rogosa-Sharpe (MRS) plates to detect the demand of D-alanine. “+” and “ and “-” represent MRS plates with and without D-alanine, respectively; Alr-COE/ΔAlr W56, pPG- −” represent MRS plates with and without D-alanine, respectively; Alr-COE/∆Alr W56, pPG-Alr- Alr-COE/ΔAlr W56; PPT/ΔAlr W56, pPG-T7g10-PPT/ΔAlr W56. COE/∆Alr W56; PPT/∆Alr W56, pPG-T7g10-PPT/∆Alr W56."
An interesting story from the @mdpiopenaccess.bsky.social archives.
First, I found this super easy #ImageForensics duplication in this 2021 Vaccines paper, doi.org/10.3390/vacc...
Shown with blue boxes.
I posted it on PubPeer, in August 2022.
pubpeer.com/publications...
/
🎥 DFRWS EU 2026 | Multimedia Forensics Join us at the Linköping University (LiU) for our new #DFRWSEU2026 Track debuting this year! 👉 Program (Friday): https://buff.ly/GwzBlHL 🧿 Register: https://buff.ly/O5gA8U6
🎥 DFRWS EU 2026 | Multimedia Forensics 🔍
Join us at the Linköping University (LiU) for our new #DFRWSEU2026 Track debuting this year!
👉 Program (Friday): buff.ly/GwzBlHL
🧿 Register: buff.ly/O5gA8U6
#MultimediaForensics #DeepfakeDetection
#ImageForensics #ComputerVision #AIGenerated
4 more retractions by @acs.org Omega for Prof. Hamad Al-Lohedan from King Saud University, totalling now 29 retractions!🏆
🤝Team work with @mumumouse2.bsky.social
@nvitiensis.bsky.social and other @pubpeer.com contributors ❤️
#ResearchIntegrity 🔍 #ImageForensics #FoSci
#ThisImageIsFine
🤝Team work with @mumumouse2.bsky.social , @nvitiensis.bsky.social and other @pubpeer.com contributors ❤️
#ResearchIntegrity 🔍 #ImageForensics #FoSci
#ThisImageIsFine
Three panels from a scientific paper, showing horizontal strips, each with a blue fluorescent, sausage-shaped core surrounded by green fluorescent clouds, and a black background. The labels suggest three different experiment, but I added orange boxes to show that the middle and bottom panel overlap with a mirroring.
A good retraction by @frontiersin.bsky.social of Neurotrope/Synaptogenix / Thomas Jefferson U paper.
The $SNPX president/CSO called me a "sir", claimed figures were just 'illustrative examples', the senior author then dropped 200 hard-to-open, un-annotated 'lsm' files - fun times.
#ImageForensics
Nine panels, showing fluorescent blue cells on a black background. Cells look circular or elliptical, some are bright, some less bright. Panels are labeled with different time points varying from 0h, 12h, 24h, 48h, 72h, 96h, and two combinations of RSV with or without Z-DEVE-FMK
Here is your #ImageForensics challenge of today.
Nine panels, different time points and different treatments. None of the panels should overlap or look the same. But.... can you spot the unexpected overlap?
This one is easy!
Screenshot of panel "B" from a scientific paper, showing grey rectangular panels in two rows, three columns, with purple blueish cells in random patterns
Can you spot the overlapping panels? Tell me which ones! The first three correct answers (either by illustration or a good description) will win that shiny Emoji Award.
It will look so good on your resume!
#ImageForensics
Screenshot from PubPeer comment: https://pubpeer.com/publications/47271D82EFC58D74875C02505E79DE "#1 Elisabeth M Bik comment accepted April 2023 Concerns about Figure 2. Spotted by ImageTwin. Cyan boxes: In the top right TEM panel, the same area appears to be visible twice, albeit rotated 180 degrees Orange boxes: Two TEM panels representing different treatments appear to overlap, albeit with a 90 degree rotation and the addition or removal of small gold particles." Below that a grid of 12 photos in three columns, four rows. I have added cyan and orange boxes to show overlapping and duplicated areas in the TEM photos (which all show small black dots on a background of cloud-like grey structures.
Fourth retraction for Surgery Professor at @lunduniversity.bsky.social :
Let's hope he is better at surgery than at electron microscopy.
pubpeer.com/publications...
#ImageForensics
#ScientificPublishing
Left: Original Figure 10, containing panels from 2 groups (A and B, in 2 columns) and 6 different organs (in 6 rows, heart, liver, spleen, stomach, lung, kidney). I have added colored boxes to highlight that several panels in the original appear to contain duplicated (photoshopped?) areas. Right: Figure in the same orientation as on the left, but all panels now look different. Still, two sets overlap
I know that @frontiersin.bsky.social are trying hard to be taken seriously.
But addressing the problems in the original figure (left) by a completely new set of panels (that also have problems; right) puts them straight back into predatory journal range.
#ImageForensics
#ScientificPublishing
/
Screenshot of a figure from a scientific paper showing mouse swimming tracks, shown as tracks in circles, 8 rows, 5 columns displaying different time points and conditions. I have labeled 5 sets with colored boxes.
How many duplicated panels in a scientific paper are still addressable with a correction? How many “oopsies” can a research group make before an Expression of Concern or retraction?
This one (anonymized) has 5 duplications in one figure and was corrected. Is that a good outcome?
#ImageForensics
Publications of Atta reported and retracted, by publisher Publisher Not Retracted || Retracted Elsevier 71 3 Wiley 28 1 MDPI 31 9 Taylor & Francis 20 0 Springer Nature 11 2 ACS 3 7 SC Virtual Company of Physics 8 0 RSC 2 3 SAGE 2 0 Others 4 0 Bentham 1 0 De Gruyter 1 0
Ayman Atta (KSU, EPRI) now boasts no less than 25 #SciRetraction! 🍾 Soon entering @retractionwatch.com 's leaderboard!🏅
🤝Team work with @mumumouse2.bsky.social @nvitiensis.bsky.social among many other @pubpeer.com
contributors ❤️
#ResearchIntegrity 🔍 #ImageForensics #FoSci
#ThisImageIsFine
Four panels in a row, 4 different labels. Each shows a light grey background with darker little dots and larger shapes (cells).
Four panels - but two overlap. Can you spot them?
#ImageForensics
Figure from a scientific paper showing four columns, each showing 11 narrowly cropped Western blot strips. Each strip has a light grey background and four darker grey or black horizontal stripes (protein bands). One panel, however, has five stripes.
One of these things is not like the other....
#ImageForensics
We see four Western blots, each of a different protein. Each blot shows three dark horizontal stripe-like bands on a lighter grey background. I have added red boxes highlighting that lanes 1 and 3 of the top blot (GAPDH) look very, very similar.
Of the 19 papers I reported in 2016, 8 have been corrected, 1 was retracted, 1 resolved, and 9 remain unaddressed.
This apparently altered photo was corrected by stating that GAPDH was run on separate gels.
How is that resolving this issue???
pubpeer.com/publications...
#ImageForensics
Figure from a scientific paper labeled Figure 1E. It shows 6 panels in 2 columns, 3 rows. Each shows white worm-like cells on a darker grey background. Two panels overlap, marked with red boxes.
Reported to @pnas.org in Feb 2016.
Asked again in Sept 2017.
Journal replied: "The authors have provided a response, and the Editorial Board is deciding how to proceed."
They are still deciding, apparently 🦗
pubpeer.com/publications...
#ImageForensics
Figure from a scientific paper showing 6 panels, labeled A through F, shown in 2 columns and 3 rows. Each panel is a square and shows liver tissues from rats, which look like pale grey/purplish background with darker blue/purple cells and some brown staining. Some arrows in each panel point towards some darker purple cells.
Spot the two overlaps and you might win an Emoji Award!
Each panel is from a differently treated rat, so there should not be any overlaps.
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
Two photos from the scientific paper, both showing an electron microscopy photo (black and white) of a white cell with lots of extrusions (tentacle like). I have added green and pink boxes to show areas that look duplicated.
An Editorial Expression of Concern by @Nature for one of Marc Tessier-Lavigne (former @Stanford president)'s papers.
pubpeer.com/publications...
www.nature.com/articles/s41...
#ImageForensics
#ScienceIntegrity
Two Western blot panels under each other (p-ERK1/2 and Tubulin). Each strip shows dark grey horizontal bands on a ligher grey background. Labels showing different time points, and different additions of EGF or Res + EGF.
It's a beauty!
Can you spot the problems here? Find at least two for your chance of winning one of the three Emoji Awards.
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
A scientific figure containing seven scanning electron microscope images of nanoparticles on nanofibers (they look kinda like brussels sprouts on stalks).
Finally got around to reporting a big batch of dodgy biomaterials articles today. This is probably the worst image of the bunch. Can you spot anything off here?
#ImageForensics
(2/X): Osman et al. 2026 (10.1016/j.ijbiomac.2025.149633) pubpeer.com/publications...
Even one overlap like this would have me question this experiment. But in Figure 6, most panels appear to overlap, supposedly representing different conditions (clearly not...).
#ImageForensics
Four Transmission Electron Microscopy photos in a 2x2 grid. Each has a light grey speckled background, and each shows darker dots in random patterns.
An easy one today.
Four TEM photos, four different formulations of 'quantum dots'.
Can you spot the unexpected overlap?
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
Figure from scientific paper, showing panels in 5 columns and 4 rows, all showing cells associated with different nanoparticles. Each panel has a black background, blue fluorescent nuclei and green fluorescent blobs.
Can you spot all three overlaps?
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
Two panel, labeled C / NA1I and D / NA1C, respectively. They show TEM (Transmission Electron Microscopy) photos, each showing a light grey background with darker black dots of different size in a sort of random pattern.
A nice and easy #ImageForensics challenge, sent to me by a reader.
Can you spot the problem here?
#ThisImageIsFine
#ScientificPublishing
Ten panels in two columns, 5 rows. Labeled A through J. Left column shows pink tissues with white crinkly lines. Right column shows brown tissue on a cyan blue background.
There are two sets of overlapping panels in this figure.
Can you spot them both?
Each panel is a differently treated rat.
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
Figure from a scientific paper showing transwell assays. We see panels in a grid of four columns and four rows. The panels show purple cells on a light grey background.
Two sets of panels look identical, and another set of panels overlaps.
Can you spot at least two of the three problems here?
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
Figure from a scientific paper. We see a grid of panels in 3 rows (different treatments) and 5 columns (different timepoints), each showing a dark grey background with lighter grey cells in small groups.
After yesterday's very hard challenge, I will give you a super easy one. Can you spot the duplication?
#ImageForensics
#ThisImageIsFine
#ScientificPublishing
No replies yet.... It is a hard one!
Hint: focus on the bottom right panel - there is something unexpected going on there.
#ThisImageIsFine
#ScientificPublishing
#ImageForensics
Figure from a scientific paper showing 9 'woundhealing' assays, where a pipette tip is used to make a scratch in a monolayer of cells and the assay monitors how fast the edges grow back together.
A pretty hard #ImageForensics.
Can you spot the duplication?
#ThisImageIsFine
#ScientificPublishing
Left: Figure 3D, showing 6 panels with orange, big and smaller blobs on a yellow background. Marked by me with rounded boxes in many different colors to highlight duplicated areas. Right: Corrected Figure 3D, showing 6 completely new panels, grey background and purple blobs.
Despite clear evidence of image manipulation, @elsevierconnect.bsky.social's BBRC did not retract the paper. Instead, authors were allowed to submit a new, “clean” set of figures.
This response seriously undermines research integrity.
#ThisImageIsFine #ImageForensics
pubpeer.com/publications...
