Improvement of collusion secure convolutional fingerprinting information codes

Abstract

This paper presents, analyzes and improves the false positive problem detected in Convolutional Fingerprinting Information Codes presented by Zhu et al. The codes are a concatenation of a convolutional code and a Boneh-Shaw code. In their paper Zhu et al. present a code construction that is not necessarily c-secure with ε-error because in their modified detection algorithm the standard Viterbi error probability analysis cannot be directly applied. In this case we show that false positives appear, that is to say, the problem of accusing an innocent and legal user of an illegal redistribution. In this paper, we give a bound on the probability of false positives and justify it analytically and by simulations. Moreover, some guidelines for a correct design of this family of codes is also given. © 2009 Springer-Verlag Berlin Heidelberg.