imagetwin

From our webinar: ▫️Policies ≠ operational ▫️Checks too late ▫️Workflows inconsistent Shift: reactive → proactive screening at submission 🗒️Webinar recap: lnkd.in/dFm5Za9T 🎥Recording: lnkd.in/dF27StaA

I continue slowly working through one volume of AJCR (w/an Associate Editor-in-Chief in Wafik S. El-Deiry, MD, PhD). So much crap. This one paper was so bad that I gave up before finishing (I’m only flagging images identified by @ImageTwinAI for reasons).

"For sale: scientific reputation": Publishers seek means to detect paper mills and "risk holding up the conveyor belt of science." horizons-mag.ch/2026/02/26/for…

Our first Imagetwin webinar on research integrity is now available to watch.🎙️💚 Key takeaway: Manual checks don’t scale as submission volumes grow. Editorial workflows need automated screening and structured integrity checks. Watch the recording 👇 tinyurl.com/mc2pmd

Imagetwin is hosting Protect Research Integrity: How to Build a Proactive Integrity Culture. Make sure to attend it on February 25. linkedin.com/events/protect…

We’re launching our Research Integrity Webinar Series with MindCrafted Analytics. 📅 Feb 25 | 14:30 CET | LinkedIn Live 👉Sign up now: linkedin.com/events/7427648…

Our new interface is live! We redesigned Imagetwin’s web app to make daily workflows clearer and more intuitive. What’s new👇 ✅ Cleaner navigation ✅ Easier batch scans across multiple documents ✅ Instant access to scan history via the History tab

C2PA AI Verification is live on Imagetwin. We now check images for Content Credentials (C2PA) to confirm AI generation using cryptographic metadata. ✅ Clear AI attribution ✅ Metadata-based verification ✅ Industry-backed standard (@OpenAI , @Adobe , @Microsoft).

Dana-Farber Cancer Institute settled a $15M False Claims Act case over misrepresented or duplicated images in National Institutes of Health grant applications (reported by @RetractionWatch). Many image issues were visible for years. Early, systematic screening changes outcomes.

✅ More accurate detection within subimages (e.g. western blots) ✅ Better identification of duplicates across papers ✅ More consistent results overall

@ImageTwinAI image integrity checks are now built into the Rivyr Publishing Suite. Editors can flag duplication, manipulation, and AI-generated visuals directly in their workflow. 🔗 Full announcement: shorturl.at/q0GrV

“AI can streamline workflows but only when it supports human judgment.” Key takeaway from @Integra_Global ’s event our @ImageTwinAI team joined. Thank you to all the speakers and organizers!

Nice to be acknowledged. "The retraction has been agreed following an investigation of the concerns raised by Penicillium citreoviride (@ImageTwinAI) and Actinopolyspora biskrensis (me) on PubPeer, which identified concerns regarding Figures 2 and 3." pubpeer.com/publications/E…

Big news! We’ve partnered with Imagetwin to bring advanced image screening into Editorial Pilot. Detect duplication, manipulation & AI visuals, screen 100M+ figures, & review 10 manuscripts seamlessly. A new standard for research integrity is here: integranxt.com/press-releases……

Integra and Imagetwin Partner to Elevate Research Integrity with Advanced Image Screening in EditorialPilot - stm-publishing.com/integra-and-im…

@kaveh1000 @addictedtoigno1 To clarify: every box marks a region our system found duplicated elsewhere in the figure. In dense cases you get many overlapping hits, so the overlay can look extreme, that’s a reflection of how much of the image is reused, not a judgment about the authors’ effort or intent.

@addictedtoigno1 To be honest the boxes make it worse than it seems. I have seen 100s of similar cases. Could be @ImageTwinAI is picking these up in a more granular way than before. Anyway here is a partial animation: youtu.be/7bkVema-AVg

Reverse Image Forensics Challenge: Try to find a unique area in Figure 9 of this recent Scientific Reports paper: 10.1038/s41598-025-17456-6 [Aug 2025] - Annotated by ImageTwin

@addictedtoigno1 Thanks for sharing this!

@Thatsregrettab1 Appreciate you highlighting this!

Our upgraded microscopy-image model is now 2× more accurate.🔍 It handles dense textures, low contrast, and crowded fields with far better consistency. ✅ Higher true-positive recall ✅ Reliable even on lower-quality samples ✅ Refined with real user feedback