Image Manipulation Detection Software in Research

Image manipulation in scientific figures can mislead readers and compromise research integrity. Our tool detects inappropriate splicing, cloning, hidden edits, and more, in research images.
Get Started

Trusted Image Integrity Solutions

Advanced forensic analysis to detect manipulations in research images and safeguard scientific credibility.

Detects Inappropriate Edits

Our AI model identifies signs of splicing, copy-move forgeries, and localised manipulation in western blot images.

Designed for Scientific Research

Trained on real-world research images, our system detects scientific image manipulation relevant to the academic publishing process.

Continuously Refined for Accuracy

Our manipulation detection is regularly improved using real-case feedback and emerging manipulation techniques.

Scientific Image Manipulation Detection

Manipulated images can distort the scientific record, whether through splicing, cloning, or contrast enhancement. Imagetwin’s manipulation detection tool helps researchers, academic institutions, and publishers flag inappropriate visual edits before publication.

Our AI model uses deep learning and forensic image analysis to identify suspect areas in western blot figures, detecting both splicing and copy-move forgeries under a unified manipulation detection model.

What Image Manipulations can Imagetwin detect?

Imagetwin is purpose-built to detect a wide range of image manipulations that may compromise the integrity of scientific figures. These include:

Splicing

Splicing involves cutting and merging parts of different images, often seen in western blots or gel electrophoresis. Imagetwin detects both vertical and horizontal splices and highlights affected regions using a colour-coded heatmap.
  • Detects both vertical and horizontal splice seams in western blots and gel images
  • Highlights affected regions with a colour-coded heatmap and confidence score
  • Apply filters and adjust heatmap colouring for detailed visual examination

Copy-Move Forgeries

Duplicating regions within an image is a common tactic to hide unwanted elements or falsely reinforce data. Imagetwin’s unified alteration model detects cloned, stamped, or resampled areas and highlights them using a colour-coded heatmap.
  • Identifies cloned or resampled regions within a single image
  • Assigns a single confidence score per image to help reviewers prioritise flagged cases

Heatmap Visualisation

Flagged alterations are displayed as a colour-coded heatmap overlaid on the original image, making it easy to identify exactly where inconsistencies appear. The detail view shows three panels: the original image, a filtered version for enhanced visibility, and the heatmap result. Users can adjust heatmap colouring and transparency.

  • Three-panel view: original image, filtered version, and colour-coded heatmap
  • Customise heatmap colouring and transparency for detailed examination

How Imagetwin Detects Image Manipulations

Imagetwin detects image manipulation in three simple steps.

1

Paper or Image Submission

User uploads entire manuscripts, files or individual images to our secure platform for analysis.

2

AI-Powered Analysis

Our manipulation detection engine performs a detailed forensic scan of each image using deep learning and pattern recognition techniques.
3

Confidence Score & Reporting

Receive a comprehensive report highlighting detected issues. Each flagged image is assigned a single confidence score, helping users to prioritise and assess risks.

Powerful Features for Manipulation Detection

AI-Powered Image Recognition

Detects splicing, copy-move forgeries, and localised manipulation in western blot images using a unified detection model.

Confidence Scores

Flagged images are assigned a detection probability score, helping users to assess risks more accurately.

Bulk Scanning

Analyze large datasets efficiently with our batch-scanning feature. Ideal for publishers, journals, and institutions.

Comprehensive Reports

Easily generate detailed PDF reports summarising all flagged manipulations with visual evidence.

Data Encryption

Your data is protected with industry-standard encryption and security best practices.

All-In-One Integrity Analysis

Our platform also detects image duplication, plagiarism and AI-generation.

Why Choose Imagetwin?

Image manipulation in scientific figures presents a serious challenge to research integrity. Imagetwin provides targeted tools to detect these issues early and accurately.
Trained on Scientific Imagery

The system is trained specifically on western blot images, allowing it to detect subtle and complex forms of manipulation including splicing and copy-move forgeries.

Our detection process is designed to be efficient, helping journals, publishers, and institutions identify concerns early in the publication process.
We prioritize data encryption and privacy, ensuring your files remain confidential.
Detection capabilities are regularly updated based on real-world use, emerging manipulation techniques, and input from academic users.
Our technology is developed with input from publishers, universities, and researchers, ensuring real-world applicability in academic and scientific publishing.

Protect Research Integrity with Confidence

Start using Imagetwin to detect image integrity issues and support trustworthy research publishing.

Get Started
Contact Us

Frequently asked questions

Imagetwin detects splicing (both vertical and horizontal), copy-move forgeries, and other localised alterations in western blot images. The system uses a unified alteration detection model that highlights inconsistent regions using a colour-coded heatmap.

Each flagged image is shown in a three-panel detail view: the original image, a filtered version with enhanced visibility, and a colour-coded heatmap highlighting regions of concern. Users can adjust heatmap colouring and transparency for detailed examination.

Currently, the dedicated manipulation detection model is applied to western blot images. Other image types such as microscopy and histology are covered by Imagetwin’s duplication and plagiarism detection features.

The current model achieves a true positive rate of approximately 83% with a false positive rate under 2%, based on testing against real-world cases from PubPeer and challenging edge cases. The model is continuously updated to improve accuracy.

Duplication detection identifies reused figures within or across manuscripts. Manipulation detection identifies alterations made within a single image, such as splicing seams or cloned regions in western blots. Both features are included in every Imagetwin scan.