The ultimate steganalysis benchmark?

We present a new benchmark for binary steganalysis meth- ods, based on the asymptotic information (in the entropic sense) it gives about the presence of hidden data. The the- oretical foundation is quite unlike ad hoc performance mea- sures found in steganalysis literature that are based on false positive and negative rates. It is argued that this new met- ric is an application-independent long-run measure of true performance. There are some challenges to computing the benchmark empirically, and some suggested methods are presented, but no definitive answer emerges. As a simple case study, some steganalysis methods from the literature are evaluated using these techniques.