Proofig AI for Publishers

Detect duplicate images in scientific papers and automate integrity checks with Proofig powered by AI.

Four microscopy images with 180-degree rotations, highlighting the need for precise analysis.

Why is image integrity important?

As one Journal of Clinical Investigation editorial clearly states: “Data accuracy forms the foundation of the scientific enterprise, and without it, the enterprise risks crumbling.” (1)

While most publishers review manuscripts to ensure they include reliable information, image integrity issues such as duplications, manipulations, reuse from published manuscripts, alterations, and AI-generated images are very difficult to detect. In fact, the percentage of manuscripts flagged for image-related problems ranges from 20 to 35 percent.

If manuscripts are published with integrity issues, it can lead to costly investigations, damaging the reputations of the publishers. The impact factors of the journals are also affected, reducing the public’s trust in scientific research.

Set of four microscopy images with scaling, highlighting concerns in data integrity

What kind of image integrity software does Proofig offer?

We’ve developed a cutting-edge AI solution for the automatic identification of scientific image issues pre and post publication. Proofig can detect:

Full duplications, including scaling or rotating
Partial duplications or overlapping
Flipped images, rotations, cloning
"Paper mill" derived fraudulent manuscript images
AI-generated images
Reuse from published manuscripts by accessing PubMed’s Source database
Any combination of these issues

Proofig is committed to continually evolving and expanding its feature set to include new models, ensuring we stay ahead in the race to protect scientific integrity.

How does Proofig benefit publishers?

Proofig provides you with a game-changing pre-publication solution for ensuring each image in each of thousands of manuscripts is reliable.

With Proofig, publishers can:

Analyze a large number of images within minutes
More easily identify fraudulent manuscripts derived from paper mills
Automate screening of articles to find image duplications, manipulations, and AI-generated images
Reduce human image identification errors
Protect a journal's reputation and impact factor
Significantly reduce time and costs associated with image proofing

All manuscripts and reports are kept confidential.

Four microscopy images with cropping and flipping, highlighting data manipulation

How does Proofig work?

Proofig AI leverages advanced machine learning techniques and an extensive database to ensure the integrity of scientific images by identifying various forms of image integrity issues.

Publishers can check each manuscript, confidentially and automatically, in four simple steps:

1. Upload Your Manuscript

Publishers begin by uploading their PDF manuscript to the Proofig platform. The system extracts all images and sub-images from the manuscript for analysis.

2. Analyze for Integrity Issues

Proofig AI analyzes the manuscript for multiple types of image issues:

Detection of Reuse from Published Manuscripts: By accessing PubMed’s Source database, which contains tens of millions of images, Proofig can identify reused sub-images, commonly known as plagiarism.
Duplication and Reuse Within a Manuscript: Proofig identifies duplication or reuse of sub-images within the same manuscript. This includes detecting instances of scaling, rotation, flipping, full and partial overlap, ensuring each image is unique.
Alteration and Manipulation: The system detects alterations or manipulations within a single sub-image, such as cloning, editing, deletion, and splicing, ensuring the authenticity of each sub-image.
Detection of AI-Generated Images: Proofig AI identifies images created by the most widely used AI model. Proofig is committed to continually evolving and expanding its feature set to include new models, ensuring we stay ahead in the race to protect scientific integrity. Take Our Quiz To Test Yourself

3. Validate and Review Results

Publishers can check the detected sub-images and validate the results. They can then review any suspected issues flagged by Proofig AI.
The system highlights suspected integrity issues, allowing publishers to add only relevant findings to the final report. Proofig also provides advanced further investigation tools to help users make well-informed decisions.

4. Generate a Comprehensive Report

Proofig AI processes the selections and produces a comprehensive report in minutes. The report includes both a page view of the suspected issue and a detailed view of each confirmed issue.
The report helps in ensuring the integrity of the scientific images before publication.

Data Privacy and Security

At Proofig, we prioritize the privacy and security of our users' data. We want to make it unequivocally clear that:

Training Data: Our system is trained exclusively using material we have developed and open-source content designated for commercial use, ensuring compliance with legal and ethical standards.
User Data: We do not use any data from user-uploaded manuscripts to train our models or algorithms. User-uploaded data is analyzed solely to detect image integrity issues within the submitted manuscript and generate the corresponding report.
Confidentiality: All analyses are conducted on private, secure servers. Your manuscripts and reports remain fully confidential and are not used for any other purposes.

By adhering to these strict data privacy and security protocols, Proofig ensures that your data is protected while maintaining the highest standards of image integrity analysis.

Proofig AI is trusted by leading publishers, research institutes, and researchers

Who uses Proofig AI already?

Proofig’s clients include integrity position holders from top universities and prominent international publishers such as Springer Nature, AACR, SAGE Publishing, RSC, JCI, Taylor & Francis. They also include PIs and senior researchers in the life sciences.

Proofig AI has been quoted in Nature and by COPE as a leader in the field of image integrity.

Where can I find out more?

Anchor 1