Efficient identification of traitors in fingerprinted multimedia contents

Abstract

In a fingerprinting scheme a different set of marks is embedded in each copy of a digital object, in order to deter illegal redistribution. A group of dishonest users, called traitors, collude to create a pirate copy that hides their identities, by putting together different parts of their copies. If the sets to be embedded are the codewords of an error correcting code then efficient algorithms can be used to trace the guilty. In this paper we present a tracing algorithm that, by taking advantage of soft-decision decoding techniques, finds all possible identifiable traitors.