Image Analysis
Common integrity issues in images
As publication volumes grow, reused figures without attribution, internal duplications, and questionable edits become harder to spot manually. Plagium helps speed up screening and prioritize what needs review.
Identify images or parts of images reused without proper attribution.
Detect duplications within the same set of figures or materials.
Flag signals of edits, crops, and alterations that require review.
Support visual authenticity review with AI-generation signals.
Fast screening, safe review
Upload images for analysis and receive a report to guide review. The ImageTwin engine compares figures and identifies similarities, even with crops or adjustments.
Select the images that require verification in your editorial or research workflow.
The analysis matches visual patterns and flags cases that deserve attention.
Focus on what matters: flagged items, context, and next steps.
A more in-depth analysis than a visual check
The analysis combines automated detection and integrity signals to flag duplication, possible manipulation, reuse across literature, and signs of synthetic image generation. The goal is faster screening and organized evidence for human review.
Reports and evidence to decide with clarity
The analysis organizes findings by risk type and helps prioritize review. Use the report for internal audits, editorial screening, and author communication.
Build visual integrity into your process
Image Analysis is designed for workflows where transparency and traceability matter, from submission intake to audits and internal reviews.
Support pre-publication screening and reduce risk before accepting a manuscript.
Strengthen integrity policies and standardize review with consistent reporting.
Catch inadvertent reuse and validate figures before submission or publication.
Examples and reports without leaving the site
You can view reports and reference materials right here. If you prefer, you can also open them in a new tab.
Use it for screening and validate with human review
Flagged findings are not, by themselves, proof of an issue. Interpret results with context, raw data, and internal policies.
