As technology advances, the number of applications relying on face recognition is
on the rise. While facial recognition technologies have many benefits, it's important
to use them in a responsible manner in order to avoid privacy risks. In this paper
we analyze the privacy risks of the processing of face imprints generated by face
recognition technologies. We characterize the risks of re-identification attacks against
facial imprint databases in multiple scenarios regarding different attacker strength.
Our findings show that even if a large number of subjects are surveilled and the attacker
can only inefficiently benefit from using the embeddings, the risk of re-identification
is still concerning.