Steganography is the art and science of invisible communication. Its aim is the transmission of information embedded invisibly into carrier data. The goal of steganalysis is to discover steganographic alterations to carrier data. The project BEST (Better Steganalysis) explored the influence of image content when detecting steganography. If this influence, which interferes with the detector statistics, is reduced, the reliability of steganalysis can be increased. Steganalysis can be regarded as an attempt to separate image (or carrier) content from embedded steganographic signals. This project consisted of two phases. During its first phase, a new technique to scan elements of potential steganograms allowed a more fine-grained evaluation of the group size, a parameter of the RS attack by Fridrich et al., which affects the detection performance. The results of Phase 2 include newly developed attacks for advanced JPEG steganography like MB2 that use targeted features to blockiness reduction artefacts. Phase 2 also comprises a new generic methodology to apply spatial domain attacks in the DCT domain. All attacks are implemented in a software package called steganographic workbench (swb), which is described in this report. The package also includes implementations of some previous state of the art attacks that have been used for evaluation.
展开▼
