Cryptography Reference
In-Depth Information
contents in order to protect intellectual properties or the ownership of the
original multimedia sources [5]. Typical watermarking schemes were based on
transform-domain techniques using discrete cosine transform (DCT) [6], dis-
crete wavelet transform (DWT) [7] and discrete Fourier transform (DFT) [8],
spatial-domain methods [9], VQ domain schemes [10, 11], for example, to em-
bed the watermark into certain coe
cients in their respective domains. These
schemes have been popular research topics in the last decade.
As we stated in Chapter 1, the three most critical requirements for de-
signing watermarking algorithms are:
watermark imperceptibility
,
watermark
robustness
,and
watermark capacity
. After considering the three fundamental
requirements for watermarking, we determine in this chapter, how to fix the
watermark capacity, and how to employ tabu search [12, 13] to find a tradeoff
between watermark imperceptibility and watermark robustness. Tabu search
is an evolutionary algorithm, a meta-heuristic approach, and is character-
ized by the use of a flexible memory. It is able to eliminate local minima
and to search areas beyond a local minimum. There are few research pa-
pers concentrating on designing watermarking algorithms with optimization
techniques [14, 15] in literature. One of the contributions in this chapter is
how to employ tabu search, an optimization technique, to obtain an optimal,
watermarked image.
Different from conventional schemes used to perform watermarking in the
spatial or transform domains, we propose in this chapter an innovative algo-
rithm concentrating on vector quantization (VQ) [16]. This is the VQ-based
image watermarking, suitable for error-resilient transmission over lossy net-
works and noisy channels. Searching research literature relating to digital
watermarking, only a few research papers concentrate on the error resilient
transmission of watermarked media [17, 18, 19, 20]. In addition, most of the
watermarking related literature focus on how to resist intentional attacks by
applying benchmarks to watermarked media in order to assess the effective-
ness of the watermarking algorithm. Therefore, we provide another viewpoint
to protect the ownership of the original, and simultaneously retain the recon-
structed image quality with error-resilient coding.
At the beginning of watermarking research in mid-1990s, researchers
examined the robustness of watermarking algorithms by imposing some
attacking schemes, including the low-pass filtering, compression, geometric
distortion, flipping, for example, into the watermarked media. These attack-
ing schemes were integrated into several benchmarks like Stirmark [21]. These
benchmarks seldom consider the scenario of transmitting the watermarked
media over the communication channels [22]. This is the motivation of the
research in this chapter. With the inspiration that there is little research re-
lating to transmission of watermarked media over the transmission channels,
we combine the two topics. That is, watermarking and error resilient coding,
and this has led to promising results. By using tabu search, we obtain an
optimized solution suitable for the transmission of watermarked images over
lossy channels.
