Synopsis of “Digital Watermarking”

Submitted to: Submitted by:

Mr. Sheo Kumar Mohit Kumar
Mr. Amit Kumar Yadav 0709010054
CS- III year

Abstract
The growth of networked multimedia systems has created the need for the
copyright protection of various digital medium, e.g., images, audio clips,
video, etc. Copyright protection involves the authentication of ownership
and the identication of illegal copies of a (possibly forged) image. One
approach used to address this problem is to add a visible or invisible
structure to an image that can be used to seal or mark it. These structures
are known as digital watermarks. The watermark is capable of carrying such
information as authentication or authorization codes, or a legend essential
for image interpretation. This capability is envisaged to find application in
image tagging, copyright enforcement, counterfeit protection, and
controlled access. In this paper, we rest outline the desirable characteristics
of digital watermarks. Previous work in digital watermarking is then
summarized. Several recent approaches that address these issues are also
discussed.

Introduction
Digital media facilitate efficient distribution, reproduction, and
manipulation over networked information systems for image, audio clips,
and videos. However, the fact that an unlimited number of perfect copies
can be illegally produced is a serious threat to the rights of content
owners.
However, these efficiencies also increase the problems associated with
copyright enforcement. A number of technologies are being developed
to provide protection from illegal copying. They include:
(1) Encryption Methods {the use of public and private keys to encode the
data so that the image can only be decoded with the required key,
(2) Site Security Methods {the use of firewalls to restrict access,
(3) Using Publicly Accessible Low Quality Thumbnail images, and
(4) Digital Watermarking , this includes the robust unobtrusive labeling
of an image with information pertaining to copyright, and the use of
image checksums or other techniques to detect the manipulation of
image data.
To address the non-obtrusive copyright enforcement issue, digital
watermarks (i.e., author signatures) are under investigation.
Watermarking is the process of encoding hidden copyright information
in an image by making small medications to its pixel content. Unlike
encryption which protects content during the transmission of the data
from the sender to receiver, digital watermarking does not restrict access
to the image information. Watermarking compliments encryption by
embedding a signal directly into the data. Thus, the goal of a watermark
is to always
Remain present in the data to provide solid proof of ownership. It should
be noted that embedded signaling or watermarking can be used for a
variety of other purposes other than copyright control. For example, it
can be used for owner identication, to identify the content owner,
fingerprinting, to identify the buyer of the content, for broadcast
monitoring to determine royalty payments, and authentication, to
determine whether the data has been altered in any manner from its
original form. However, here we restrict our discussion here to issues that
are related to copyright control. Although there are two main divisions of
watermarks, e.g., visible and invisible, this paper focuses on algorithms
and techniques for invisible watermarks. In general, there are two basic
requirements of invisible watermarks. The watermarks should be (1)
perceptually invisible and (2) Robust to common signal processing and
intentional attacks. Early research on digital watermarking concentrates
on the rest objective without considering the second one. Recently much
work has been devoted to designing robust watermarking schemes.
Perceptual models have also been incorporated to make the best trade
between perceptual invisibility and robustness to signal processing .
The goal of this paper is to give a brief summary of various digital
watermarking techniques available for the purpose of authentication,
forgery detection, and copyright enforcement. The paper is organized as
follows. In the next section, we outline desirable properties of a
watermark for copyright control, which can be quite different from
watermarks for authentication purposes.

SPECIFICATIONs of Watermarking Techniques

To be effective, the watermark should be: (1) perceptually invisible within

the host media; (2) statistically invisible to thwart unauthorized removal;
(3) readily extracted by the image owner and (4) robust to accidental and
intended signal distortions incurred by the host image, e.g., filtering,
compression, re-sampling, re-touching, cropping, etc.
Unobtrusive (Difficult to notice)
The watermark should be perceptually invisible to the viewer nor should
the watermark degrade the quality of the content. In earlier work ,Cox.,
had used the term “imperceptible", and this is certainly the ideal.
However, if a signal is truly imperceptible, then perceptually based lossy
compression algorithms should, in principle, remove such a signal.
Current state-of-the-art compression algorithms probably still leave
room for an imperceptible signal to be inserted. This may not be true of
next generation compression algorithms. Thus, to survive the next
generation of lossy compression algorithms, it will probably be necessary
for a watermark to be noticeable to a trained observer.
Robustness
The watermark must be difficult (hopefully impossible) to remove. Of
course, in theory, any watermark may be removed with sufficient
knowledge of the process of insertion. However, if only partial
knowledge is available, for example, the exact location of the watermark
within an image is unknown, then attempts to remove or destroy a
watermark by say, adding noise, should result in severe degradation in
data fidelity before the watermark is lost. In particular, the watermark
should be robust to the following attacks and characteristics. Universality
The same digital watermark algorithm should apply to all three media
types. This is potentially helpful in the watermarking of multimedia
products. Also, this feature is conducive to implementation of audio,
image, and video watermarking algorithms on common hardware.
Tamper-resistance the watermarking techniques should be robust to
legitimate signal distortions as well as intentional attacks to remove or
tamper with the digital watermark.
Common Signal Processing The watermark should still be retrievable
even if common signal processing operations are applied to the data.
These include, digital-to-analog and analog-to-digital conversion,
resampling, requantization (including dithering and recompression), and
common signal enhancements to image contrast and color, or audio bass
and treble, for example. Common Geometric Distortions Watermarks in
image and video data should also be immune from geometric image
operations such as rotation, translation, cropping, and scaling.
Subterfuge Attacks: Collusion and Forgery In addition, the watermark
should be robust to collusion by multiple individuals who each possess a
watermarked copy of the data. That is, the watermark should be robust to
combining copies of the same data set to destroy the watermarks.
Further, if a digital watermark is to be used as evidence in a court of law,
it must not be possible for colluders to combine their images to generate
a different valid watermark with the intention of framing a third-party.

A Framework for Watermarking

A watermarking framework consists of three parts:

(1) The watermark,
(2) The marking algorithm, and
(3) The verification algorithm.
Each owner has a unique watermark which the owner would like to
embed into his/her proprietary work. The marking algorithm
incorporates the watermark into
The multimedia medium. The verification algorithm authenticates the
watermarked information, determining both the owner and the integrity
of the image. There are currently numerous techniques for applying a
digital watermark to an image. The techniques can be divided into two
major categories based on the desired application for the watermark (1)
to detect image tampering and (2) to embed copyright information. The
techniques used to detect image tampering tend to be fragile and
introduce insignificant data loss. Robust watermark algorithms used to
embed copyright data tend to introduce increased visible artifacts, the
notable exception are the spread spectrum methods of digital
watermarking which are particularly useful for copyright labeling, being
both robust and invisible.
Further classification of digital watermarking can be achieved by
categorizing the image data by the robustness of the watermarking
technique and the obtrusiveness, the amount of visible artifacts (or data
loss) introduced. With the limited scope of this paper, we will focus only
on digital watermarking algorithms that embed copyright information
into the targeted media.
There are several major algorithm areas with many variations. One major
algorithm is based on the modification of Least Significant Bit (LSB) of
the pixel content .Watermarks also can modify the images spectral or
transform coefficients directly. These algorithms often modulate Discrete
Cosine Transform (DCT) coefficients according to a sequence known only
to the owner. Watermarking techniques may be image dependent. These
techniques increase the level of the watermark in the image while
maintaining the imperceptibility of the mark .For example, one of the
wavelet methods incorporates features from most of the above
techniques. Its implementation lends itself to watermarking data rate-
scalable video. Time stamps thwart a clever attack proposed by IBM on
all of these watermarking schemes. Visible watermarks also exist; IBM has
developed a proprietary visible watermark to protect images that are
part of the digital Vatican library project. The watermarking itself is only
a small part of any controlled access and distribution scheme; a method
for secure distribution would combine encryption with digital
watermarking. Lastly there is the hybrid technique where many of these
techniques may be used in combination with each other. The sections
below describe these watermarking algorithms in detail.

Digital Watermarking Techniques

Least Significant Bit Modification

The most common and early watermarking approaches modify the least
significant bits (LSB) of an image based on the assumption that the LSB data are
insignificant. Two LSB techniques are described in. The author first replaces the
LSB of the image with a pseudo-noise (PN) sequence, while the second adds a
PN sequence to the LSB of the data. And another early watermarking method
obtains a checksum of the image data, and then embeds the checksum into the
LSB of randomly chosen pixels. Others add a modified maximal length linear
shift register sequence to the pixel data. They identify the watermark by using
the spatial cross-correlation function of the modified sequence and part of the
watermarked image. The Digimarc Corporation describes a method that adds
or subtracts small random quantities from each pixel. Addition or subtraction is
determined by comparing a binary mask of L bits with the LSB of each pixel. If
the LSB is equal to the corresponding mask bit, then the random quantity is
added, otherwise it is subtracted. The watermark is subtracted by first
computing the difference between the original and watermarked images and
then by examining the sign of the difference, pixel by pixel, to determine if it
corresponds to the original sequence of additions and subtractions. The
Digimarc method does not make use of perceptual relevance and is probably
equivalent to adding high frequency noise to the image. As such, it may not be
robust to low-passltering. Turner proposed a method for inserting an
identification string into a digital audio signal by substituting the “insignificant"
bits of randomly selected audio samples with the bits of an identification code.
Bits are deemed “insignificant" if their alteration is inaudible. Such a system is
also appropriate for two dimensional data such as images, as discussed in.
Unfortunately, Turner's method may easily be circumvented. For example, if it is
known that the algorithm only an effect the least significant two bits of a word,
then it is possible to randomly ip all such bits, thereby destroying any existing
identification code.
In a recent paper, Macq and Quisquater briefly discuss the issue of
watermarking digital images as part of a general survey on cryptography and
digital television. The authors provide a description of a procedure to insert a
watermark into the LSB of pixels located in the vicinity of image contours. Since
it relies on modifications of the least significant bits, the watermark is easily
destroyed. Further, their method is restricted to images, in that it seeks to insert
the watermark into image regions that lie on the edge of contours.
Information Tagging
Caronni suggests adding tags {small geometric patterns - to digitized images at
brightness levels that are imperceptible. While the idea of hiding a spatial
watermark in an image is fundamentally sound, this scheme is susceptible to
attack by filtering and redigitalization. The fainter such watermarks are the
more susceptible they are such attacks and geometric shapes provide only a
limited alphabet with which to encode information. Moreover, the scheme is
not applicable to audio data and may not be robust to common geometric
distortions, especially cropping. Brassil et al. propose three methods
appropriate for document images in which text is common. Digital watermarks
are coded by: (1) vertical shifting text lines, (2) horizontally shifting words, and
(3) altering text features such as the vertical end lines of individual characters.
Unfortunately, all three proposals are easily defeated, as discussed by the
authors. Moreover, these techniques are restricted exclusively to images
containing text.

Quantization Noise Embedding

Tanaka et al. describe several watermarking schemes that rely on embedding
watermarks that resemble quantization noise. Their ideas hinge on the notion
that quantization noise is typically imperceptible to viewers. Their first scheme
injects a watermark into an image by using a predetermined data stream to
guide level selection in a predictive quantizer. The data stream is chosen so that
the resulting image looks like quantization noise. A variation on this scheme is
also presented, where a watermark in the form of a dithering matrix is used to
dither an image in a certain way. There are several drawbacks to these schemes.
The most important is that they are susceptible to signal processing, especially
requantization, and geometric attacks such as cropping. Furthermore, they
degrade an image in the same way that predictive coding and dithering can.

Statistical Techniques
Bender describes two watermarking schemes. The first is a statistical method
called”Patchwork" that somewhat resembles the statistical component of Cox's
proposal. Patchwork randomly chooses n pairs of image points, (ai; bi), and
increases the brightness at ai by one unit while correspondingly decreasing the
brightness of bi. The expected value of the sum of the differences of the n pairs
of points is then claimed to be 2n, provided certain statistical properties of the
image are true. In particular, it is assumed that all brightness levels are equally
likely, that is, intensities are uniformly distributed. However, in practice, this is
very uncommon. Moreover, the scheme may (1) not be robust to randomly
jittering the intensity levels by a single unit, and (2) be extremely sensitive to
geometric affine transformations. The second method is called “texture block
coding", wherein a region of random texture pattern found in the image is
copied to an area of the image with similar texture. Autocorrelation is then used
to recover each texture region. The most significant problem with this
technique is that it is only appropriate for images that possess large areas of
random texture. The technique could not be used on images of text, for
example. Nor is there a direct analogy for audio.

Frequency Spectrum-Based Methods

Koch proposes two general methods for watermarking images. The first method
breaks up an image into 8 8 blocks and computes the Discrete Cosine Transform
(DCT) of each of these blocks. A pseudorandom subset of the blocks is chosen,
then, in each such block, a triple of frequencies is selected from one of 18
predetermined triples and modified so that their relative strengths encode a 1
or 0 value. The 18 possible triples are composed by selection of three out of
eight predetermined frequencies within the 8 8 DCT block. The choice of the 8
frequencies to be altered within the DCT block is based on a belief that the
“middle frequencies... have moderate variance", i.e., they have similar
magnitude. This property is needed in order to allow the relative strength of the
frequency triples to be altered without requiring a modification that would be
perceptually noticeable. Superficially, this scheme is similar to our own proposal
and, in fact, also draws analogy with spread spectrum communication.
However, the structure of their watermark is different from ours. The set of
frequencies is not chosen based on any perceptual significance or relative
energy considerations. Further, because the variance between the eight
frequency coefficients is small, one would expect that their technique may be
sensitive to noise or distortions. This is supported by the experimental results
which report that the “embedded labels are robust against JPEG compression
for a quality factor as low as about 50%". An earlier proposal by Koch and Zhao
used not triples of frequencies but pairs of frequencies, and was again designed
specifically for robustness to JPEG compression. Nevertheless, they state that “a
lower quality factor will increase the likelihood that the changes necessary to
superimpose the embedded code on the signal will be noticeably visible". In a
second method, designed for black and white images, no frequency transform is
employed. Instead, the selected blocks are modified so that the relative
frequency of white and black pixels encodes the final value. Both watermarking
procedures are particularly vulnerable to multiple document attacks. To protect
against this, Koch and Zhao propose a distributed 88 created by randomly
sampling 64 pixels from the image. However, the resulting DCT has no
relationship to that of the true image and consequently may be likely to cause
noticeable artifacts in the image and be sensitive to noise.

Checksum Technique
This watermark is formed from the checksum value of the seven most significant
bits of all pixels. A checksum is the modulo-2 addition of a sequence of fixed-
length binary words. It is a special type of hash function. In this technique, one
word is the concatenation of eight 7-bit segments, which come from eight
different pixels. Each pixel is involved in the checksum only once. The final
checksum is fifty-six bits. The technique then randomly chooses the locations of
the pixels that are to contain one bit of the checksum. The pixel locations of the
checksum, together with the checksum itself, form the watermark. The last bit
of each chosen pixel is changed (if necessary) to equal the corresponding
checksum bit. This value must be kept secret. To verify this watermark the
checksum of a test image is obtained, and compared to the ideal version in
watermark. Any discrepancy invalidates the image. The advantages of this
technique are: (1) the embedding watermark only changes (on average) half of
the pixels covered by watermark; (2) an image may hold many watermarks as
long as they do not overlap, and (3) this method is very fast. On the other hand,
the disadvantages of this technique are: (1) this watermarking method is fragile.
Any change to either the image data itself or the embedded checksum can
cause the verification procedure to fail, (2) the checksum method does not
detect pixels swaps or similar attacks. A forger could replace a section with one
of equal size and checksum, and (3) an attacker could remove the entire
watermark by replacing the LSB plane.

Hybrid and Other Techniques

The hybrid approach combines several techniques together to synthesize a new
variation of the watermarking algorithm. For example, Walton uses a checksum
on the image data which is embedded in the least significant bits of certain
pixels. Others add a maximal length linear shift register sequence to the pixel
data and identify the watermark by computing the spatial cross-correlation
function of the sequence and the watermarked image. Watermarks can be
image dependent, using independent visual channels, or be generated by
modulating JPEG coefficients. These watermarks are designed to be invisible, or
to blend in with natural camera or scanner noise. Visible watermarks also exist;
IBM has developed a proprietary visible watermark to protect images that are
part of the digital Vatican library project. In addition to direct work on
watermarking images, there are several works of interest in related areas.
Adelson describes a technique for embedding digital information in an analog
signal for the purpose of inserting digital data into an analog TV signal. The
analog signal is quantized into one of two disjoint ranges, (0, 2, 4..., 1, 3, 5..., for
example) which is selected based on the binary digit to be transmitted. Thus
Adelson's method is equivalent to watermark schemes that encode information
into the least significant bits of the data or its transform coefficients. Adelson
recognizes that the method is susceptible to noise and therefore proposes an
alternative scheme wherein a 2 1 Hadamard transform of the digitized analog
signal is taken.
The differential coefficient of the Hadamard transform is o
set by 0 or 1 unit prior to computing the inverse transform. This corresponds to
encoding the watermark into the least significant bit of the differential
coefficient of the Hadamard transform. It is not clear that this approach would
demonstrate enhanced resilience to noise. Furthermore, like all such least
significant bit schemes, an attacker can eliminate the watermark by
randomization.

Conclusion
The proliferation of network multimedia systems dictates the need for
copyright protection of digital property. To conclude, any successful
watermarking algorithm would have to exploit properties of the human
visual system and combine these with effective modulation and
channel coding. Future work will concentrate on producing watermarks
that are robust to filtering, lossy image compression, noise corruption
and changes in contrast. In addition these algorithms must anticipate
possible attacks on the integrity and security of the watermark and to
devise suitable countermeasures. This paper serves as a brief summary
on several more recent and popular digital watermarking techniques
for multimedia information systems.