Authentication codes and watermarking for multimedia communications in lossy networks

The security of lossy multimedia communications in the Internet motivates a soft multimedia authentication system that tolerates acceptable data manipulations. This work is concerned with the development and use of soft Message Authentication codes (MACs) and digital watermarking to provide data integrity mechanisms that “bend” rather than “break” when a received multimedia signal is slightly different from what was sent.; The first approach towards a soft authentication system is to develop soft MACs. Several soft image MAC algorithms are proposed in this work. A new Approximate Authentication Code (AMAC) is developed and it serves as an approximate cryptographic primitive. Unlike conventional MACs, the AMACs are probabilistic in nature and have the property that they change slowly as the Hamming distance between the original and modified data increases. Based on the AMAC, the approximate Image Authentication Codes (IMACs) are designed using techniques such as block averaging and smoothing, parallel AMAC bit computation, and image histogram enhancement. The AMACs and the approximate IMACs are both based on the bit representations of the image. Another feature-based IMAC, the histogram based image MAC (H-IMAC), is developed by bringing spatial information in the histogram representations and calculating the checksum bits for each dimension.; Digital watermarking techniques are used as the second half of the solution. By digital watermarking, the MACs are not attached but embedded in the image. A new authentication watermarking method for still images and its extension to MPEG2 video are developed and implemented. A novel edge based video MAC is proposed towards video authentication.; Theoretical analysis is performed on the probabilistic properties of the proposed MAC schemes. Their robustness to image compression and the capability to detect image tampering are investigated. Simulations results are presented to validate the analysis. The probability of watermark detection is studied and the simulation results follow the theoretical expectation.