Insertion, Detection, and Extraction of Messages Hidden by Optimal Multi-Signature Spread Spectrum Means
Abstract
We carry out optimized spread-spectrum data embedding in a given digital image (or audio, or video sequence). First, the overall image/host medium is pre-processed into transform-domain small blocks from which host vectors are obtained via zig-zag scanning vectorization. Multiuser data embedding is performed in the generated host vectors. Under this data embedding model, we calculate an orthogonal set of embedding spread-spectrum signatures that achieves maximum sum signal-to-interference-plus noise ratio (sum-SINR) at the output of the linear-filter receivers for any fixed embedding amplitude values. Then, for any given total embedding distortion constraint, we present the optimal multi-signature assignment and amplitude allocation that maximizes the sum capacity of the embedding procedure. The practical implication of the reported results is sum-SINR, sum-capacity optimal multiuser/multi-signature spread-spectrum data embedding in the digital medium. Extensive experiments that we carried out demonstrate the effectiveness of the proposed methods. The problem of extracting blindly data embedded over a wide band in a spectrum (transform) domain of a digital medium is also considered.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 09, 2015
- Accession Number
- ADA621874
Entities
People
- Dimitris A. Pados
Organizations
- University at Buffalo