Coding Theorems on the Simple Capacity for Digital Fingerprinting Codes

摘要

Digital fingerprinting codes provide a method to protect licensed digital contents against illegal redistribution. We focus on the situation where two malicious users collude and generate an overwritten codeword by using a memoryless collusion channel. We investigate coding theorems for the digital fingerprinting codes using a simple decoder such that each user is judged as malicious or innocent from an overwritten codeword together with his/her codeword. We first discuss the case where the collusion channel is known. We give a formula of the simple capacity. Next, we consider the case where the collusion channel is unknown. We succeed in giving the formula of a simple capacity of this case as well.