The fuzzy commitment scheme is a cryptographic primitive that can be used to store biometric templates being encoded as fixed-length feature vectors protected. If multiple related records generated from the same biometric instance can be intercepted,
their correspondence can be determined using the decodability attack. In 2011, Kelkboom et al. proposed to pass the feature vectors through a record-specific but public permutation process in order to prevent this attack. In this paper, it is shown that this countermeasure enables another attack also analyzed by Simoens et al. in 2009 which can even ease an adversary to fully break two related records. The attack may only be feasible if the protected feature vectors have a reasonably small Hamming distance; yet, implementations and security analyses must account for this risk. This paper furthermore discusses that by means of a public transformation, the attack cannot be prevented in a binary fuzzy commitment scheme based on linear codes. Fortunately, such transformations can be generated for the non-binary case. In order to still be able to protect binary feature vectors, one may consider to use the improved fuzzy vault scheme by Dodis et al. which may be secured against linkability attacks using observations made by Merkle and Tams.
