Collusion-secure fingerprinting and B/sub 2/-sequences

Abstract

We discuss a strategy initiated by Boneh and Shaw (see IEEE Trans. Inform. Theory, vol.44, p.1897-1905, 1998) for collusion-secure fingerprinting. We show that under this strategy, finding fingerprinting schemes that resist coalitions of two users amounts to finding B/sub 2/-sequences of binary vectors. A sequence of vectors v/sub 1/, v/sub 2/,..., v/sub n/ is a B/sub 2/-sequence if all sums v/sub i/+v/sub j/, 1/spl les/i/spl les/j/spl les/n, are different: the associated extremal set-theoretic problem is what is the maximal size of a B/sub 2/-sequence? We shed new light on this old combinatorial problem and improve on previously known upper bounds.