Capacity-achieving fingerprint decoding

We study randomized fingerprinting codes that achieve the fundamental capacity limits subject to the so-called Boneh-Shaw marking assumption. Two decoding schemes are studied in particular: the joint decoder is capacity-achieving but computationally intense, while the simple decoder is suboptimal but efficient. We provide tight bounds as well as numerical results for capacities and study the difference between these two schemes. Finally, security strategies for both the fingerprint embedders and the collusive attackers are presented.