#ImageForensics
LightNews LightNews
promptfoundry.bsky.social
DINO-FIL uses multi-layer DINOv3 ViT-H+/16 features with a lightweight fusion decoder for pixel-level localization of generative inpainting forgeries. It generalizes across four of five cross-dataset…

#DINOv3 #ForgeryLocalization #ImageForensics #ComputerVision
https://arxiv.org/abs/2610.21780
October 18, 2026 at 2:01 AM
futuregearai.bsky.social
New research proposes the first continual learning approach for IFL, since current methods localize manipulations well but struggle to adapt to newly emerging forgeries. Tests of existing IFL and continual learning…

#ImageForensics #ContinualLearning #AIResearch
https://arxiv.org/abs/2609.38251
October 1, 2026 at 12:01 PM
tomk.eurosky.social
Well done, @mumumouse2.bsky.social and @nvitiensis.bsky.social ! 🤝
Retraction #41 for JRX from Saveetha engineering college, India
#ResearchIntegrity 🔍 #ImageForensics #FoSci #ThisImageIsFine 🕵️ #ScientificIntegrity #Academia
spottingthespot.bsky.social spottingthespot.bsky.social @spottingthespot.bsky.social · Sep 4
And one more for Joseph Raj Xavier! This time @tomk.eurosky.social didn't cause the retraction...

doi.org/10.1007/s117...
September 5, 2026 at 4:44 PM
academic-integrity.bsky.social
Al-Gorair et al. 2023 (DOI: 10.1016/j.ijoes.2023.100245), published in Elsevier's International Journal of Electrochemical Science taking full advantage of his photoshop license! pubpeer.com/publications...
#imageforensics
August 16, 2026 at 10:48 PM
reeserichardson.bsky.social
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
A scientific figure containing seven scanning electron microscope images of nanoparticles on nanofibers (they look kinda like brussels sprouts on stalks).
January 17, 2026 at 10:43 PM
reeserichardson.bsky.social
Several within-image duplications in a 2018 Acta Biomateriala (Elsevier) article, cited 88 times, originating from Harvard Medical School.

#ResearchIntegrity #ImageForensics
Fluorescence images of cell cultures where duplicated regions have been highlighted by colored boxes.
March 15, 2025 at 5:12 PM
reeserichardson.bsky.social
Some within-image duplications in a 2015 Journal of Materials Science (Springer Nature) article, cited 33 times.

Author response: "...any perceived duplication likely arises from natural structural similarities within the sample."

#ImageForensics #PublicationIntegrity #AcademicIntegrity
An image from a scanning electron microscope with some unexpected duplications, highlighted by colored boxes.
March 12, 2025 at 4:06 PM
reeserichardson.bsky.social
Copious duplications in cell culture images. 2020 article in Elsevier journal Polymer Testing, cited 131 times. Authors from Iran, Canada, Malaysia and Norway. #AcademicIntegrity #PublicationIntegrity #ImageForensics
Red and yellow image of cells. Some duplicated regions are indicated with boxes of various colors. Red and yellow image of cells. Some duplicated regions are indicated with boxes of various colors. Red and yellow image of cells. Some duplicated regions are indicated with boxes of various colors.
March 10, 2025 at 3:03 PM
reeserichardson.bsky.social
I have officially commented on more than 1,000 comments on PubPeer! 🥳

@pubpeer.com

#ScienceIntegrity #ImageForensics
PubPeer search for users:"Reese Richardson" that returns 1008 results.
February 6, 2025 at 8:32 PM
reeserichardson.bsky.social
A desperately-needed win today: an article I reported to the publisher in August 2024 has been retracted.

pubpeer.com/publications...

Four problematic articles by the same authors still remain in this journal. Some more dupes:

pubpeer.com/publications...

#ResearchIntegrity #ImageForensics
Duplications found in Article 1. Image showing duplicated particles common to three articles by the same authors across three journals.
January 22, 2025 at 6:41 PM
jennifermach.bsky.social
Holy cats this paper I’m working on had so much image reuse with different labels. Won’t see the light of day now because we got it post acceptance but prior to publication. #SciPub #ImageForensics
May 19, 2025 at 9:07 PM
jennifermach.bsky.social
If I were a PI, I would not submit any paper without running it through an integrity analysis––most of the #ImageForensics issues we find seem to be honest mistakes with file handling.
May 13, 2025 at 12:17 PM
hansonmark.bsky.social
#PhDChat #SciPub #ResearchIntegrity #ImageForensics #OpenScience #OpenAccess #ScientificPublishing #AcademicSky #Science #ScienceFunding #SaveTheNIH
August 4, 2025 at 11:27 AM
hansonmark.bsky.social
Voila! The Research Integrity Feed hashtag list for post inclusion 📊🔍👀

#ResearchIntegrity
#PredatoryPublisher
#PredatoryPublishing
#EditorialIndependence
#SciRetraction
#PaperMill
#PaperMills
#HijackedJournal
#HijackedJournals
#ImageForensics
#FoSci
#ThisImageIsFine
#DoctoredImage

Cheers 🍻
January 5, 2025 at 10:41 PM
hansonmark.bsky.social
Welcome! Added #ImageForensics ! And I've now also added #ThisImageIsFine

Fun world I'm learning of. I'll update the tagged 1st post with the full list of hashtags once the list settles. But right now have many more included.
January 4, 2025 at 7:07 PM
hansonmark.bsky.social
Nice. Too many to include in the description, but these ones below are all entered now. Will update 1st post with list later.

#ResearchIntegrity
#PredatoryPublisher
#PredatoryPublishing
#EditorialIndependence
#SciRetraction
#PaperMill
#FoSci
#HijackedJournal
#DoctoredImage
#ImageForensics
January 4, 2025 at 6:33 PM
curlyneuron.bsky.social
🐬✖️ What connects the X-Files and dolphins? Apparently, #ScientificIntegrity! 🤔

Read all about it here: www.theexperimentalist.org.

Special thanks to Dr. @elisabethbik.bsky.social for this interview❣️

#ImageForensics
The Experimentalist
"Scientists are often quick to blame the media for poor #scicomm , but ultimately those stories wouldn’t exist without the scientists participating in the process" Let's change that. Talk abou...
www.theexperimentalist.org
December 17, 2025 at 4:24 PM
ocasapiens.bsky.social
#Nettascienza

Premiati elisabethbik.bsky.social, ormai con più medaglie del maresciallo Zhukov, e PubPeer@mastodon.social che le querele-bavaglio non riescono a censurare.
November 18, 2024 at 12:44 PM
rossbiology.bsky.social
These gel or blot image panels struck me as having too homogenous a background. Adjusting levels (R) reveals there is very little noise in the backgrounds. Also suspicious to me is a little bit of white halo around some bands (could it be compression artifact?) What do you think? #ImageForensics
July 5, 2026 at 9:55 PM
rossbiology.bsky.social
Saw this on Facebook yesterday. I was skeptical. Some of the comments mentioned this is a different part of the sky (not in Bootes) or that the void isn't that large, etc. But that's not what I was worried about. You can't always believe what you see! #imageforensics
Post from Facebook showing "a void in space 1 billion light-years across" with numerous starts surrounding a dark (starless) area The same image but I've indicated places where parts of the image have been duplicated
March 25, 2025 at 2:48 PM
paperstarsorg.bsky.social
Look, if you aren’t aware of @elisabethbik.bsky.social yet, we don’t know what to tell you. Go follow her now. What a total badass! #imageforensics #accountability #pubpeer
June 20, 2025 at 10:17 AM
c0nc0rdance.bsky.social
🏆 You get the Bluesky ImageForensics Award! Cut and paste as needed.

I'm using this source: https://emojipedia.org/trophy/
July 26, 2023 at 9:27 PM
c0nc0rdance.bsky.social
Glad you're here!

People use feeds here instead of hashtags, and they're more formal: you have to whitelist people into them. I'm still figuring all this out, but you should start an ImageForensics feed.

Look at the satellite dish icon "My Feeds" to see what it looks like.
July 26, 2023 at 11:57 AM
colinpurrington.bsky.social
Here's a less pixelated one in case anyone wants a look. I was hoping it would be easier to see overlaps but I could only see the one already mentioned.
Post Elisabeth 'unethical nutjob' Bik ‪@elisabethbik.bsky.social‬ #ImageForensics Spot the overlapping panels. Figure from a scientific paper, showing panels in 6 columns (Week 1, Week 4, Week 20, PCL/PU 4) and 4 rows (H&E, MT, TB, CD 68). Each panel shows skin tissue sections, stained pink (H&E), or shades of brown, purple, green and blue. ALT February 3, 2025 at 3:06 PM Everybody can reply 1 quote 12 likes Write your reply ‪Stacey Ogden 🦔‬ ‪@labogden.bsky.social‬ · 29m CD68 PU week1 and week 4. Week 4 looks zoomed and offset from week 1. ‪Solal Pirelli‬ ‪@solalpirelli.bsky.social‬ · 32m PU4/CD68 left of week 1 and right of week 4 are the same, I think? ‪Edgar Häner‬ ‪@gaussianwalk.bsky.social‬ · 35m Cannot see it, looking forward to the answer :-) Colin Purrington ‪@colinpurrington.bsky.social‬ Home Search Notifications Chat Feeds Lists Profile Settings New Post Search Following Popular With Friends Mentions Mutuals Gardening OnlyPosts The 'Gram What's Cooking 🍽️ Fungi Friends iNaturalist More feeds Trending Rep. Raskin Starlink Starmer Critique Beyoncé Bruno and Gaga Kanye West Feedback • Privacy • Terms • Help Figure from a scientific paper, showing panels in 6 columns (Week 1, Week 4, Week 20, PCL/PU 4) and 4 rows (H&E, MT, TB, CD 68). Each panel shows skin tissue sections, stained pink (H&E), or shades of brown, purple, green and blue. Figure from a scientific paper, showing panels in 6 columns (Week 1, Week 4, Week 20, PCL/PU 4) and 4 rows (H&E, MT, TB, CD 68). Each panel shows skin tissue sections, stained pink (H&E), or shades of brown, purple, green and blue.
February 3, 2025 at 8:44 PM
ikashnitsky.phd
+ #DoctoredImages #ImageForensics
cc @elisabethbik.bsky.social @charlespiller.bsky.social
January 4, 2025 at 3:09 PM