App repackaging is a common threat in the Android ecosystem. To face this threat, the literature now includes a large body of work proposing approaches for identifying repackaged apps. Unfortunately, although most research involves pairwise similarity comparison to distinguish repackaged apps from their "original" counterparts, no work has considered the threat to validity of not being able to discover the true original apps. We provide in this paper preliminary insights of an investigation into the Multi-Generation Repackaging Hypothesis: is the original in a repackaging process the outcome of a previous repackaging process? Leveraging the Androzoo dataset of over 5 million Android apps, we validate this hypothesis in the wild, calling upon the community to take this threat into account in new solutions for repackaged app detection.
展开▼
