Amped Authenticate Update 13901: Introducing Projects and Report, PRNU with Multiple Reference Cameras, Improved Social Media Identification, and More

The extra-hot summer we’ve had in Italy didn’t stop our dev team! So here we are again this week with another release, this time for Amped Authenticate, maybe just in time for you to try it before your summer holidays!

Projects and Report

We are proud to introduce this widely requested feature: you have now the ability to bookmark results, add comments, organize bookmarks in folders, and finally generate a fully-detailed Report for your case. Below is how Authenticate’s new interface will look like at the end of an image investigation:

The Project panel provides an intuitive way for the user to organize the work: whenever the user finds an interesting result, they can bookmark it using the Project panel. An Amped Authenticate project is made of a list of bookmarks, possibly organized into folders.

When a bookmark is added, the currently selected filter (including its input and post-processing parameters), together with the path to the currently loaded evidence and reference images are saved into an entry in the Project panel.

After adding a bookmark, the user can change the name, set a description and a status either directly from the Project panel, or opening the bookmark editing dialog shown below.

The bookmark description is a free text, useful for commenting what is shown, while the bookmark status is meant to quickly resume the positive or negative contribution of the bookmarked result towards image integrity or authenticity. Three status values are possible: “Ok”, “To be checked”, and “Warning”, which will be reflected in the color of the square next to the bookmark name (green, yellow and red, respectively). Of course, you can leave the status unset (gray color).

Folders can be added by right-clicking on the panel and selecting “Add folder”, so to group bookmarks for better organization of your work.

When you finish, you can click on Tools -> Generate Report to obtain a detailed forensic report which provides general project information, software version, and for each bookmark: the associated evidence and reference files together with their MD5 hash, a description of the filter and the values set for all its parameters, bibliographic references, and of course the filter output! This is how the Report’s table of contents looks like:

And this is a sample Report entry for the ADJPEG filter:

When saving the project to a .aap file, all paths to your referenced files will be made relative to the project’s destination folder. And when loading a project, the MD5 hash value of all files will be checked to ensure they did not change.

We are confident this feature will make your working experience an Authentic – ate joy!

PRNU Identification with Multiple Reference Cameras

While Authenticate already allowed checking multiple images in a folder against the currently loaded Camera Reference Pattern (CRP), if you wanted to test multiple CRPs against some images you had to iteratively load them each time, manually. This is not the case anymore!

When you load a CRP file, its name will be added to the list of configurations of the PRNU Identification filter. This way, you can load several CRPs once and check the current images against each of them. If the currently loaded image has negative compatibility with a CRP, its name will be written in red (warning).

This means that you can, of course, click on different CRPs and see the filter output update accordingly but, more interestingly, it means you can run a Batch Processing and let Authenticate check all images in a folder against all currently loaded configurations (so-called “N-to-M comparison”).

If you’ve looked carefully at the image above, you may have noticed there’s one additional new entry in the PRNU Identification filter, let’s zoom in a bit on that:

Since we’re dealing with multiple CRPs, we believed it was necessary to add the CRP MD5 hash value in the filter output, so you’ll always be able to link a filter result to the exact CRP file used to generate it. This MD5 value will go in the Report as well. For maximum security, the hash is re-computed every time you match an image to the CRP. You can disable MD5 computation from the Program Option panel if you need to save time when processing lots of pictures.

Improved and Simplified Social Media Identification Filter

The ability to detect whether an image was processed by a Social Media Platform is crucial in our crazy-for-sharing world. Authenticate’s Social Media Identification filter was based on a k-NN (k Nearest Neighbor) classifier to compare the “feature descriptor” extracted from your image to thousands of descriptors available in the internal database. If at least three sufficiently close elements (from the same SMP) were found, Authenticate classified the file as coming from that SMP. We decided to simplify the filter and make it more “safe”:

  • now a match is returned only when an identical descriptor is found in the database (we don’t require image resolution to be identical);
  • the filter interface has been simplified by removing the “Date of reference file” (because it was a confusing element for some users and it did not add much to the analysis), and the “Similarity Score” (because it no longer makes sense: if a match is returned, the score is always 1.0);
  • the messages produced by the filter in the “Result” row have been clarified: you’ll either get “No compatible Social Media Platform signature(s) found in the DB” or “Image is compatible with Social Media Platform signature(s) in the DB”, to make it clear that we are only able to draw conclusions based on Authenticate’s internal or user-defined databases, which of course could be incomplete.

Keep in mind that it may happen, like before, that your image matches multiple SMPs, as in the example below: it means that we found at least one file from each of those SMPs whose descriptor is identical to your image’s descriptor. You can ask our support team to provide the reference file for including it in your report (just tell us its Id, row 4).

Totally Renewed Samples Folder

We love providing samples for our users to play with. Now that Authenticate features Projects, we renewed the whole samples folder (accessible by clicking on Help -> Open samples folder), which now provides 15 examples. Each example is accompanied by a project file showing how to walk through it and, when needed, by additional files that are useful for the sample workflow.

File Format: Improved Quantization Table Related Warning Messages

During training sessions, we realized that the warning messages produced by the File Format filter when the input image’s JPEG Quantization Table was not found in Authenticate’s DB, could be misleading for some. It is important to understand that the fact that an image’s JPEG QT is not present in the database does not mean that the image integrity is broken: we may just miss that QT! To make this clearer, we changed the warning messages to:

  • No compression signature available for this camera in the DB: you’ll get this message when Authenticate’s DB does not contain any compression signature for the device model declared in your image metadata;
  • Compression signature not found in the DB for this camera: you’ll get this message when Authenticate’s DB contains at least one compression signature for the declared device, but none of the available signatures match your image’s one.

Other New Features

  • Batch File Format Comparison: now shows the MD5 value computed on each file, so to allow spotting duplicates (just click on this row’s name to sort by its values).
  • JPEG QT: We expanded the internal database of Quantization Tables with more than 250 new signatures. The list of compatible cameras is now sorted alphabetically, and so is the list of signatures available in the internal database. Moreover, when a user adds a new QT to its database, the filter output is immediately refreshed to include the new signature.
  • Social Media Identification: We expanded the internal database of SMPs compression signatures with images recently shared through Facebook, Whatsapp, Tumblr, Flickr, Instagram.
  • PRNU Tampering: the output map has been improved, green borders due to padding are no longer present.

Bug Fixes

We always strive to identify any bugs before we release an update, but sometimes they slip through the net! Thanks to reports sent in by you, we are able to release fixes like these below:

  • JPEG QT: fixed a bug that prevented restoring the currently loaded Reference image after running “Find image with same QT”
  • JPEG QT: fixed a bug that caused “Find image with same QT” to show the Evidence file twice in the output table.
  • File Format: fixed a bug that prevented the translation of warning messages.
  • PRNU Identification: fixed a bug that caused miss-detection of square images when rotated by 90 or 270 degrees.
  • GUI: fixed a bug that caused the viewer panel to keep showing the output map even when the currently selected configuration was deleted.
  • GUI: fixed a bug that prevented keyboard shortcuts from working as expected.
  • GUI: fixed a bug that prevented user-defined configurations in the Filter panel to be properly saved to file.
  • GUI: fixed a bug that prevented the “Reprocess current configuration” button from working as expected.

Don’t Delay – Update Today

If you have an active support plan you can update straight away by going into the menu “Help” > “Check for Updates Online” within Amped Authenticate. If you need to renew your SMS plan, please contact us or one of our authorized partners.

Table of Contents

Share on

Subscribe to our Blog

Related posts

How to Amped FIVE

Video Deinterlacing

Good day, jolly people! We are back this week with our series “Learn and solve it with Amped FIVE”  to talk about filter tips for

Read More »