Public watermark detection using multiple proxies and secret sharing

Abstract

A central issue in public watermarking schemes is the design of a detector that will not reveal sufficient information that leads to the erasure of embedded watermark, even if an adversary knows the detection algorithm and the public detection key (if any). Insofar, there is no such detector in "stand-alone" setting that achieves satisfactory security requirements. Recently, [1] gives a zero-knowledge detector that achieves security by introducing a server. We propose an alternative setting of public watermarking that involves multiple servers. In this setting, the owner keeps a secret watermark W. A verifier, given an image J (or any digital media), wants to detect whether J is watermarked. The detection is to be carried out by a group of independent proxies. The owner does not trust the verifier nor any individual proxy, thus wants to keep W hidden from them. On the other hand, the verifier does not trust the owner and any individual proxy either, and wants to protect himself against cheating. The proxies, as a group, are tasked to maintain the secrecy of the watermark, and protect the interest of the verifier. We propose a scheme based on secret sharing schemes which support arithmetic operations. The security is maintained if not too many individuals (including the proxies, the owner and the verifier) collude. The proposed scheme is efficient in terms of computation cost, and the number of rounds and bandwidth required in the communications. The scheme is arguably easy to implement. © Springer-Verlag 2004.