Date of Award
Doctor of Philosophy in Electrical Engineering - (Ph.D.)
Electrical and Computer Engineering
Ali N. Akansu
Richard A. Haddad
Yun Q. Shi
Nasir D. Memon
This dissertation presents an in-depth study of oblivious data hiding with the emphasis on quantization based schemes. Three main issues are specifically addressed:
1. Theoretical and practical aspects of embedder-detector design.
2. Performance evaluation, and analysis of performance vs. complexity tradeoffs.
3. Some application specific implementations.
A communications framework based on channel adaptive encoding and channel independent decoding is proposed and interpreted in terms of oblivious data hiding problem. The duality between the suggested encoding-decoding scheme and practical embedding-detection schemes are examined. With this perspective, a formal treatment of the "processing" employed in quantization based hiding methods is presented. In accordance with these results, the key aspects of embedder-detector design problem for practical methods are laid out, and various embedding-detection schemes are compared in terms of probability of error, normalized correlation, and hiding rate performance merits assuming AWGN attack scenarios and using mean squared error distortion measure.
The performance-complexity tradeoffs available for large and small embedding signal size (availability of high bandwidth and limitation of low bandwidth) cases are examined and some novel insights are offered. A new codeword generation scheme is proposed to enhance the performance of low-bandwidth applications. Embeddingdetection schemes are devised for watermarking application of data hiding, where robustness against the attacks is the main concern rather than the hiding rate or payload. In particular, cropping-resampling and lossy compression types of noninvertible attacks are considered in this dissertation work.
Sencar, Husrev T., "Oblivious data hiding : a practical approach" (2003). Dissertations. 611.