Collusion Resistant Multimedia Fingerprints

Abstract

Multimedia fingerprint is a technique for investigating the distribution of multimedia content (such as audio, image, video, etc.) and securing them from unapproved reallocation. Unique approval information is inserted or embedded into each scattered copies of multimedia data, with each multimedia content fingerprint allocated to a proposed recipient. Collusion attack is a well-known attack on multimedia fingerprinting; by which attackers combine multiple copies of the same multimedia content, but with dissimilar fingerprints. Therefore collusion poses a real challenge to protect multimedia information. This thesis surveys the state of the art collusion resistant digital fingerprinting algorithms and presents an analysis of these fingerprinting techniques. We investigate elementary linear and nonlinear collusions on self-determining fingerprints and analyze severities. In this thesis, we developed a web application implementing collusion resistant fingerprint algorithms and study their performance under collusion attacks.