Characterization of steganographic algorithms using software metrics

The inclusion of data hiding techniques in everything from consumer electronics to military systems is becoming more commonplace. This has resulted is a growing interest in benchmarks for embedding algorithms, which until now has focused primarily on the theoretical and product oriented aspects of algorithms (such as PSNR) rather than the factors that are often imposed by the system (e.g., size, execution speed, complexity). This paper takes an initial took at these latter issues through the application of some simple and well known software engineering metrics: McCabe Complexity and Halstead Software Measures. This paper illustrates an approach that applies these metrics to create a hypothetical, language-independent representation of an algorithm, identifying the encapsulated, measurable components that compose that algorithm. This is the first step in developing a representation that will not only allow for comparison between disparate algorithms, but describe and define algorithms in such a way as to remove language and platform dependency. Bringing these concepts to their logical conclusion highlights how such an approach would provide existing benchmarking systems a more in-depth and fair analysis of algorithms in the context of systems as a whole, and decrease variability which affects the accuracy of the theoretical and product measures used today.