14 February 2020 | Fawrensian

Major Improvements to Fawrensian

Now that our website is up and running, we’re happy to share with you the release of the first update of our forensic solution for image forgery detection: Fawrensian. We’re proud to announce major improvements to our initial version of Fawrensian (i.e., v0.5, see below) by enriching the information we provide and including a number of new functionalities.

A snapshot of Fawrensian v0.5

Among the new features available in Fawrensian v1.1 are:

  • Improved GUI: We’ve completely redesigned the graphical interface of Fawrensian to improve user experience and provide more information about each image under analysis.
  • Reliability indicators: We’ve added reliability indicators (TPR and FAR) that can aid users in establishing confidence and robustness levels for their target applications.
  • Bounding boxes to highlight forgeries: We now use bounding boxes to signal the detected forgeries across the image content.
  • Smarter analysis of saturated regions: In case of a positive double compression detection, the user can filter out saturated regions that were already present in the genuine content of the image to avoid false positives due to saturation.

With these new features, this is the new look of Fawrensian v1.1:

A snapshot of Fawrensian v1.1

Sleek, eh?

In addition, Fawrensian v1.1 allows the analysis of images embedded in PDF files, and since not only inconsistent JPEG compression traces can be found in JPEG files, also uncompressed image file formats are supported, such as, BMP, PNG, and TIFF, among others. Finally, the core algorithm of Fawrensian has also been improved, so that a higher detection performance is achieved and a more appealing visual representation of the forgery map is provided.

Stay tuned! There is more to come in Fawrensian v1.2. Any feedback from your side is welcome.

Don’t forget that the visualization of the results and/or other features can be adapted to your needs. Contact us and let us know ↗