Chapter3

TCRMERCADO | USeP – IC | 2010 1

DR. TAMARA CHER R. MERCADOUniversity of Southeastern Philippines

Institute of Computing

《 Models of Watermarking 》


Contents

3.1 Communications3.2 Communication-based models of

watermarking3.3 Geometric models of watermarking3.4 Basics of Digital Image3.5 Image Watermarking Example


Components of Communication System

m: the message we want to transmitx: the codeword encoded by the channel encodern: the additive random noise y: the received signalmn: the received message

Fig. 3.1 Standard model of a communication system

3.1 》 Communications


Components of Communication System

• encodersource coder:

modulator:

maps a message into a sequence of symbols drawn from some alphabet.

converts a sequence of symbols into a physical signal that can travel over the channel.

• the transmission channel is assumed noisy, thus an additive noise n is added to the original signal x during transmission.

• decoder receives signal y ( x + n ), inverts the encoding process and attempts to correct transmission errors.



Secure Transmission

• Passive adversary: passively monitors the transmission channel and attempts to illicitly read the message

• Active adversary: actively tries to either disable the communication or transmit unauthorized messages

• Two defense approaches: Cryptography and Spread Spectrum



Cryptography

Fig. 3.2 Standard model of a communication channel with encryption

• Prior to transmission, cryptography is used to encrypt a message using a key.

• The encrypted message (ciphertext) is transmitted over the channel• At the receiver, the ciphertext is received and decrypted using the related

key to reveal the cleartext

3.1 》 Communications 》 Secure Transmission


Cryptography

Fig. 3.2 Standard model of a communication channel with encryption

• Two uses of cryptography:• prevent passive attacks in the form of unauthorized reading of the

message.• prevent active attacks in the form of unauthorized writing.

• Downside:• does not necessarily prevent an adversary from knowing that a

message is being transmitted.• provides no protection against an adversary intent on jamming or

removing a message before it can be delivered to the receiver.



Spread Spectrum

Encodingkey

Decodingkey

Fig. 3.3 Standard model of a communication channel with key-based channel coding

• Against signal jamming (the deliberate effort by an adversary to inhibit communication between two or more people)

• Modulation is done according to a secret code, which spreads the signal over a wider bandwidth than required

• Frequency hopping - One of the earliest and simplest spread spectrum technologies



Cryptography vs. Spread Spectrum

• Spread spectrum communications and cryptography are complementary.• Spread spectrum guarantees delivery of signals. Cryptography

guarantees secrecy of messages. It is thus common for both technologies to be used together.• Spread spectrum can be thought of as responsible for the

transport layer, and cryptography as responsible for the messaging layer.



Communication and Watermarking3.2 》 Communication-Based Models of Watermarking

• Watermarking is, in essence, a form of communication where we communicate a message from the watermark embedder to the watermark receiver.

• Ways to incorporate the cover Work into the traditional communications model

1. The cover Work is considered purely as noise (Basic Model).2. The cover Work is still considered noise, but this noise is provided to the

channel encoder as side information.3. Cover Work is not considered as noise, but rather as a second message

that must be transmitted along with the watermark message in a form of multiplexing.


Informed Detector 3.2 》 Communication-Based Models of Watermarking 》 Basic Model

Fig. 3.4 Watermarking system with a simple informed detector mapped into communications model(wa: Added pattern, Co: Original cover work, cw: watermarked version of the work, cwn: noisy watermarked work)

• Watermarking is viewed as a transmission channel through which the watermark message is communicated. The cover work is part of that channel.

• Detection consists of two steps:1. Co is subtracted from the received Work, cwn, to obtain a received noisy

watermark pattern, wn. 2. wn is then decoded by a watermark decoder, with a watermark key.

• Because the addition of the coverWork in the embedder is exactly cancelled out by its subtraction in the detector, the only difference between wa and wn is caused by the noise process.

Watermark Embedder

Watermark Detector


Blind Detector

Fig. 3.5 Watermarking system with blind detector mapped into communications model. (Note that in this figure there is no meaningful distinction between the watermark detector

and the watermark decoder.)

3.2 》 Communication-Based Models of Watermarking 》 Basic Model

Watermark Embedder

Watermark Detector

• The un-watermarked cover Work is unknown, and therefore cannot be removed prior to decoding

• The received, watermarked Work, cwn, is now viewed as a corrupted version of the added pattern, wa, and the entire watermark detector is viewed as the channel decoder.


Applications3.2 》 Communication-Based Models of Watermarking 》 Basic Model

• Informed and Blind Detector models can be applied in transaction tracking or copy control, as it requires maximum likelihood that the detected message is identical to the embedded one.

• In authentication applications, the goal is not to communicate a message but to learn whether and how a Work has been modified since a watermark was embedded. For this reason, Informed and Blind Detector models are not typically used to study authentication systems.


Side Information at the Transmitter

3.2 》 Communication-Based Models of Watermarking 》 Side Information

Fig. 3.6. Watermarking as communications with side information at the transmitter.

• Much more effective embedding algorithms can be made if we allow the watermark encoder to examine co before encoding the added pattern wa .•A model of watermarking that allows wa to be dependent on co.

• The model is almost identical to Blind Detector, with the only difference being that co is provided as an additional input to the watermark encoder.•Allows the embedder to set cw to any desired value by simply letting wa = cw − co


Multiplexed Communications

3.2 》 Communication-Based Models of Watermarking 》 Multiplexed Communications

Fig. 3.7. Watermarking as simultaneous communications of two messages. (Pictured with a blind watermark detector. An informed detector would receive the original cover Work as additional input.)

• Cover Work as a second message to be transmitted along with the watermark message in the same signal, cw.

• The two messages, co and m, will be detected and decoded by two very different receivers: a human being and a watermark detector, respectively.

• The watermark embedder combines m and co into a single signal, cw.


Geometric Models of Watermarking

• Media space: a high-dimensional space in which each point corresponds to one work.

• Marking space: projections or distortions of media space.

• A watermarking system can be viewed in terms of various regions and probability distribution in media or marking space.

3.3 》 Geometric Models of Watermarking


Graphic/Image File Formats

• Graphic/Image Data Structures – Pixels : picture elements in digital images – Image Resolution : number of pixels in a digital

image (Higher resolution always yields better quality.)

– Bit-Map : a representation for the graphic/image data in the same manner as they are stored in video memory.



Geometric Models of Watermarking


• Distribution of unwatermarked works: how likely each work is

• Region of acceptable fidelity: a region in which all works appear essentially identical to a given cover work

• Detection region: describes the behavior of the detection algorithm

• Embedding distribution or embedding region: describes the effect of an embedding algorithm

• Distortion distribution: indicates how works are likely to be distorted during normal usage


Distributions and Regions in Media Space

• Works can be thought of as points in an N-dimensional media space.

• The dimensionality of media space, N, is the number of samples used to represent each work. e.g., in the case of gray scale images, this is simply the number of pixels.

3.2 》 Geometric Models of Watermarking 》 Distributions and Regions in Media Space


Types of Digital Image

• Binary Image– Each pixel is stored as a single bit (0 or 1) – A 512×512 monochrome image requires 32.768 kB

of storage.

3.4 》 Basics of Digital Image

1 0 0 1 0 1 0 0 0 1 1 0 1 0 0 10 1 1 1 1 0 0 11 1 0 0 0 1 0 1



• Gray-scale Images– Each pixel is a shade of gray, from 0 (black) to 255

(white). This range means that each pixel can be represented by eight bits, or exactly one byte.

– A 512×512 grayscale image requires 262.14 kB of storage.


138 201 90 128111 345 95 200112 122 112 7821 198 56 90

1 0 0 0 1 0 1 0



• True Color or RGB (Red-Green-Blue)– Each pixel has a color described by the amount of

red, green and blue in it.– Has a total of 256x256x256 = 16,777,216 different

possible colors in the image– 24 bit images: total number of bits required for

each pixel.– A 640×480 24-bit color image would require 921.6

kB of storage




• True Color or RGB (Red-Green-Blue)




• Indexed– Each pixel has a value which does not give its color

(as for an RGB image), but an index to the color in a color map.

– Color map or color palette is associated with the image which is simply a list of all the colous used in that image.

– Compuserve GIF allows only 256 colors or fewer in each image and so its index values only requires one byte each.




• Indexed


Pixels labeled 5 correspond to 0.2627 0.2588 0.2549, which is a dark grayish color.


The LSB Technique

• LSB: Least Significant Bit– Considered as the simplest technique for watermark

insertion.– For a 24-bit image, each pixel has 3 bytes and each

color (RGB) has 1 byte or 8 bits in which the intensity of that color can be specified on a scale of 0 to 255.

– A bright purple in color would have full intensities of red and blue, but no green. This pixel can be shown as

X0 = {R=255, G=0, B=255}– Now let’s have a look at another pixel:

X1 = {R=255, G=0, B=254}

3.5 》 Image Watermarking Example


The LSB Technique

• Since this difference does not matter much, when we replace the color intensity information in the LSB with watermarking information, the image will still look the same to the naked eye.

• Thus, for every pixel of 3 bytes (24 bits), we can hide 3 bits of watermarking information, in the LSBs.

• A simple algorithm for this technique would be:

Let W be watermarking informationFor every pixel in the image, XiDo Loop:

Store the next bit from W in the LSB position of Xi [red] byteStore the next bit from W in the LSB position of Xi [green] byteStore the next bit from W in the LSB position of Xi [blue] byte

End Loop



The LSB Technique3.5 》 Image Watermarking Example

W = TAMMY01010100 01000001 01001101 01001101 01011001

49 – 110001 11000064 – 1000000 100000166 – 1000010 100001055 – 110111 11011176 – 1001100 100110080 – 1010000 101000156 – 111000 11100082 – 1010010 1010010


The LSB Technique3.5 》 Image Watermarking Example

49 – 110001 110000 (48)64 – 1000000 1000001 (65)66 – 1000010 1000010 (66)55 – 110111 110111 (55)76 – 1001100 1001100 (76)80 – 1010000 1010001 (81)56 – 111000 111000 (56)82 – 1010010 1010010 (82)

48 55 56 65 76 82 66 81


The LSB Technique

• Watermark Extraction– take all the data in the LSBs of the color bytes and

combine them.• This technique of watermarking is invisible, as

changes are made to the LSB only, but is not robust.

• Image manipulations, such as resampling, rotation, format conversions and cropping, will in most cases result in the watermark information being lost.


Chapter3

Documents

communication channel

watermark message

watermarking watermarking

transmission channel

received message

form of communication

received work

watermarking system