Project Feature Extraction

Content-Based Authentication of Images under Format Migration

Project Participants: Veronika Krauß, Arefeh Bahrami

Supervisors: Eik List, Prof Stefan Lucks, Jakob Wenzel

Project Documentation:

  • PDF (Version: May 04 2016)
  • bib

Abstract: Archiving digital objects possesses a special kind of challenge: Thanks to digital obsolescence of both, hard- and software, archived data have to be flexible and need to stay susceptible to manipulation to guarantee access and readability in the far future. Therefore it is important to preserve the content or “value” of an object as well as a set of significant properties in order to guarantee the objects reliability.

In our project, rather than just following the regular praxis of conserving meta data, we want to find a tool, which is able to automatically determine the value of an object. Additionally we want to conserve that value in a save and unforgeable manner. We describe our search for suitable tools as well as a list of properties we consider to be important with respect to the object’s value. On the basis of possible attack scenarios we want to test the reliability of those tools; in addition we want to examine if an object created with them is still authentic. Asides from existing tools, it is also conceivable for us to develop one on our own.