ForBild

Illegal images, among other things, are being searched for during the forensic investigation of databases. First and foremost this applies to child pornography. Until now the identification of such images is based on either visual inspection or on cryptographic hashing. In the latter, an image’s hash value is being calculated and compared to the hash values stored in a database. Cryptographic hashing, however, always carries the risk of not being able to identify respective material. For example, this is always the case when file copies are not identical to the original but have undergone a format conversion.

Within ForBild, robust hashing is to be examined for its suitability to detect images in forensic investigations automatically. In contrast to the currently available solutions such as Perkeo or EnCase, these hashing processes have the characteristic that they can recognize images even after format conversions or other similar slight changes. At the same time, however, they do not respond to other, only similar-looking images. In this, they combine the principles of cryptographic hashing with the robustness against common transformations.

An appropriate integrated solution for the support of the police work can greatly extend the possibilities of current applications. The market for this therefore rests with all national and international authorities that are engaged in tracking illegal images. The application is being developed by LSK, consortium manager and longtime capture system supplier of optical storage systems to the police, and its research partners Fraunhofer SIT and TU Darmstadt. Robust hashing is being optimized and integrated for usage, while both the process of preservation of evidence is secured and the privacy aspect is observed during image capture with an innovative technical solution as well.