A Complexity-Theoretic Foundation for Digital Image Watermarking Systems

Santa Clara University, Santa Clara CA

Investigators

Abstract

Digital watermarking is being investigated as an underlying technology for digital rights management systems (which provide content authentication, copy and access control, or proof of ownership). This proposal aims to develop a complexity-theoretic foundation that allows rigorous and quantitative studies of security, transparency, efficiency and capacity of digital image watermarking systems. Unlike previous works on modeling watermarking, two fundamental constraints on the insertion and extraction algorithms are taken into account, namely, they must be fast and also preserve the perceptual fidelity of their inputs. A key effort of this research will be to obtain formal characterizations of the class of algorithms which preserve perceptual fidelity of input images under various distortion metrics based on current theories of vision. Another goal is to settle the current debate as to whether computationally secure digital image watermarking systems exist, and if so, how to design them. Other planned activities include characterizations of tradeoff relationships between transparency, security, and capacity; applications to approximate data matching problems and indexing/querying multimedia databases and libraries; and development of an undergraduate-level course in digital watermarking at Santa Clara University.