an introduction to basic concepts and watermarking and ...yao/ee4414/memon_f05_v2.pdfauthentication...
TRANSCRIPT
![Page 1: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/1.jpg)
Information Hiding, Digital Watermarking and Steganography
An Introduction to Basic Concepts and Techniques
Nasir MemonPolytechnic University, Brooklyn
![Page 2: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/2.jpg)
Information Hiding
� Information Hiding: Communication of information by embedding it in and retrieving it from other digital data.
� Depending on application we may need process to be imperceptible, robust, secure. Etc.
encoderinformationto embed
original data
retrievedinformation
markeddata
decoderprocessing
receiveddata
processed data
![Page 3: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/3.jpg)
Where can we hide?
� Media» Video» Audio» Still Images» Documents
� Software� Hardware designs� Etc.� We focus on data hiding in media. � We mainly use images but techniques and concepts
can be suitably generalized to other media.
![Page 4: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/4.jpg)
Why Hide?
� Because you want to protect it from malicious use» Copy protection and deterrence - Digital Watermarks
� Because you do not want any one to even know about its existence» Covert communication – Steganography
� Because it is ugly» Media bridging,» Meta Data embedding
� To get a free ride» Hybrid digital analog communication, captioning.
![Page 5: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/5.jpg)
Fundamental Issues
� FidelityThe degree of perceptual degradation due to embedding operation.
� RobustnessThe level of immunity against all forms of manipulation (intentional and non-intentional attacks).
� PayloadThe amount of message signal that can be reliably embedded and extracted (subject to perceptual constraints at the designated level of robustness).
� SecurityPerhaps the most misunderstood and ignored issue. Meaning
of security depends on the application as we shall see later.
![Page 6: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/6.jpg)
Classification Basis for Information Hiding Methods
� The nature of host signal, i.e., audio, video, image, text, programs, etc.
� Robust, fragile, semi-fragile. � The need for host signal at message extraction–blind or
non-blind (private or public).� The type of communication–synchronous or
asynchronous.� The threat model – intentional (malicious) or non-
intentional attacks. » Digital Watermarking» Steganography» Data Hiding
![Page 7: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/7.jpg)
Truly Interdisciplinary
� Information Theory and Communication� Signal Processing and Transforms� Game Theory � Coding Theory � Detection and Estimation Theory� Cryptography and Protocol Design
![Page 8: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/8.jpg)
A Communication Perspective
Information hiding differs from traditional communication systems in the operation of the combiner. (Beyond modulation…)
» i.e., in classical communications no “similarity” constraint is imposed on the carrier signal and its modulated version.
Communication model for Information Hiding
������������������������������������
���� ���������� ���������� ���������� ������
��������������������������������������������������������������������������������� ����� ����� ����� ����
Host Signal
Extracted Message
Signal
Message Signal
![Page 9: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/9.jpg)
Generic Model & Terminology
ChannelChannelEmbedder E
Embedder E
DetectorD
DetectorD
Host (Cover) Signal
Embedded Signal
MessageModulatorMessage
ModulatorKey
SC Y
m
W
k
^
Distorted (Attacked) Signal Message
Message
Perceptual Constraints: C ≈≈≈≈ S ≈≈≈≈ YRobustness Constraint: m=D(S)=D(Y)
Embedding: S = E (C,W)Detection: ��=D (Y,C) �� ��=D (����)
Payload: Entropy of message m
Public/Private Key (Secure Channel)
m
![Page 10: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/10.jpg)
The Parallel Between Communication and Information Hiding Systems
Embedding distortion to attack distortion (WNR)
Signal to noise ratio (SNR)
Perceptual distortion LimitsPower constraintsAttackChannel noiseHost signalSide informationEmbedder-DetectorEncoder-Decoder
Information Hiding FrameworkCommunication Framework
������������� ���������������� ����� �����
����������������������
![Page 11: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/11.jpg)
Lattice Quantization Case� Example for
» Embedding a binary symbol by use of a two-dimensional lattice.
» Embedding two binary symbols by use of two unidimensional lattices.
o → m=0× → m=1
2ℜ∈C
![Page 12: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/12.jpg)
Optimum Embedder/Detector Design
� Nested lattice codes provide an efficient algebraic structured binning scheme.» A high dimensional fine lattice is partitioned into
cosets of coarse lattices.» Embedding is by quantizing C to the nearest
lattice point in the coarse lattice.» Detection is by quantizing Y to the nearest lattice
point in the fine lattice.» The embedding rate is designated by diluting the
coset density in the fine lattice.
![Page 13: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/13.jpg)
Optimum Embedder/Detector Design
� High dimensional constructions are not feasible.
� Lattices with simpler structures are employed.» Cartesian products of low-dimensional lattices.» Recursive quantization procedures.
– Trellis coded quantization
» Error correction codes.
![Page 14: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/14.jpg)
Digital Watermarks
![Page 15: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/15.jpg)
What is a Watermark?
� A watermark is a “secret message” that is embedded into a “cover message”.
� Usually, only the knowledge of a secret key allows us to extract the watermark.
� Has a mathematical property that allows us to argue that its presence is the result of deliberate actions.
� Effectiveness of a watermark is a function of its» Stealth» Resilience» Capacity
![Page 16: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/16.jpg)
Why Watermark?
� Ownership assertion.� Fingerprinting.� Content labeling.� Copy prevention or control.� Content protection (visible watermarks).� Authentication.� Media Bridging� Broadcast Monitoring� Etc.
![Page 17: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/17.jpg)
Ownership AssertionPublic-Private Key Pair, Digital Certificate
Alice
Bob
Watermarked content
Watermark
Private Key
Original ContentJudge
Illegal copy
![Page 18: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/18.jpg)
Fingerprinting
Bob
…
…
…
Illegal copies reveal Bob’s ID
Fingerprint 1 Fingerprint 2 Fingerprint n
Illegal copies
Copy 1 Copy 2 Copy n
Original
Content
Judge
![Page 19: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/19.jpg)
Copy Prevention and ControlOriginal Content
Compliant Recorder
Bob
Content withcopy preventionwatermark
Recorder disallows more than n copies
Compliant Player
![Page 20: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/20.jpg)
Requirements
� Requirements vary with application.» Perceptually transparent - should not
perceptually degrade original content.» Robust - survive accidental or malicious attempts
at removal.» Oblivious or Non-oblivious - Recoverable with or
without access to original.» Capacity – Number of watermark bits embedded» Efficient encoding and/or decoding.
![Page 21: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/21.jpg)
Requirements are Inter-related
Perceptual Transparency
Oblivious vs. Non-Oblivious
RobustnessPayload Security
![Page 22: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/22.jpg)
Watermarking Encoding
EWatermark S
User Key K
EK(I,S)=ISource Image I
Watermarked Image IEncoder
![Page 23: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/23.jpg)
Watermarking Decoding
D
User Key K DK(J)= T
Source Image I
Decoder
δ
Watermark T
Watermark S
Yes
No
Dec
isio
n
DK(I,J,S)= {0,1}
Watermarked Image J
![Page 24: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/24.jpg)
Classification
� According to method of insertion» Additive» Quantize and replace
� According to domain of insertion» Transform domain» Spatial domain
� According to method of detection» Private - requires original image» Public (or oblivious) - does not require original
� According to type» Robust - survives image manipulation» Fragile - detects manipulation (authentication)
![Page 25: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/25.jpg)
Robust Watermarks
|
![Page 26: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/26.jpg)
Fragile Watermarks
![Page 27: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/27.jpg)
Example of a Simple Spatial Domain Robust Technique
� Pseudo-randomly (based on secret key) select n pairs of pixels . The expected value of is 0.
� Increase by 1 and decrease by 1. The expected value of is 2n.
� To detect watermark, check
( )ii ba ,
( )ia ( )ib
(((( ))))∑∑∑∑====
−−−−n
iii ba
1
(((( ))))∑∑∑∑====
−−−−n
iii ba
1
(((( ))))∑∑∑∑====
−−−−n
iii ba
1
![Page 28: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/28.jpg)
Example
![Page 29: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/29.jpg)
Additive Watermarks
W (x,y): Pseudo Random Pattern {-1,0,1}
kMultiply by gain
factor k
I(x,y) IW(x,y)
IW(x,y)= I(x,y)+ k ⋅ W(x,y)
W(x,y) detectedT< →
' ( , ) ( , )WI x y W x yR T> →
No W(x,y) detected
![Page 30: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/30.jpg)
Additive Transform Domain Technique
� Embed sequence pseudo-randomly chosen iid Gaussian samples si into perceptually significant frequency components of I (eg 1000 midband DCT coefficients).
� Example insertion formula.� To detect s in J compute.� Confidence measure is.
� Watermark remarkably robust.
iii sff αααα++++===='
( )If
( ) ( )IfJft iii −=
∑∑
∑=22
ii
iii
st
stc
iii sff α+='
![Page 31: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/31.jpg)
Example
Original
+ =
Watermark Watermarkedimage
![Page 32: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/32.jpg)
Multimedia Authentication
![Page 33: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/33.jpg)
Authentication Codes
� Provides means for ensuring integrity of message
� Independent of secrecy - in fact sometimes secrecy may be undesirable!
![Page 34: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/34.jpg)
Public-Key Cryptosystems
� Public-key cryptography was invented in 1976 by Diffie and Hellman in order to solve the key management problem. The system consists of two keys:» A public key, which is published and can be used
to encrypt messages.» A private key, which is kept secret and is used to
decrypt messages.� Since the private key is never transmitted or
shared, the problem of key management is greatly reduced.
![Page 35: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/35.jpg)
Public-Key Cryptography
The most popular public-key encryption in use today is the RSA (Rivest-Shamir-Adleman) system.
100001010010100100100101110101010010101010010100001………….
Original message
EncryptedMessage
Encryption Decryption
11100001010010100100100111110101010011101010010110001……….
Original message
Public Key Private Key
![Page 36: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/36.jpg)
Public-Key Cryptosystems for Authentication
� Certain public-key cryptographic systems in which the roles of the public and private keys in encryption and decryption can be reversed, can also be used for authentication:» Prior to sending a message, the sender encrypts the
message with his/her private key.» The message can be decrypted by the public using the
public key of the signatory (no secrecy involved).» Since it is computationally infeasible to find the private key
from the public key and the known message, the decryption of the message into meaningful text constitutes its authentication.
![Page 37: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/37.jpg)
One-Way Hash Functions
� Hash function: A computation that takes a variable-size input and returns a fixed-size digital string as output, called the hash value.
� One-way hash function: A hash function that is hard or impossible to invert, also called a message digest function.
� The one-way hash value can be thought of as the digital fingerprint of an image because:» It is extremely unlikely for two different images to hash to the
same value.» It is computationally infeasible to find an image that hashes to
a given value: precludes an attacker from replacing the original image with an altered image.
![Page 38: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/38.jpg)
One-Way Hash Functions
� Examples of hash functions used for digital signatures are:» 20-byte secure hash algorithm (SHA-1) that has been
standardized for government applications.» 16-byte MD2, MD4, or MD5 developed by Rivest.
11100001010010111010010…….0100111101010100101011101010010100001100010100………
Original Image
10001010000101000010
Image HashHashing Function
![Page 39: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/39.jpg)
111000010100101110100100100101001111010101001010111010100101000010001…………….
Original Image
10001010000101000010
Image HashHashing Function 10001010010...
Encryption
DigitalSignature
Private Key
Digital Signature Generation
� A digital signature is created in two steps: » A fingerprint of the image is created by using a one-way
hash function; » The hash value is encrypted with the private key of a public-
key cryptosystem. Forging this signature without knowing the private key is computationally infeasible.
![Page 40: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/40.jpg)
Digital Signature Verification11001011100001010010010000101001001001011000011111010101001010001010001.………….
10001010000101000010
Decryption
DigitalSignature
10001010000101000010 =?Hash of the
original imageHash of the
image in question
Hashing function
VerificationSoftware
Public Key
Image being authenticated
Yes or No
![Page 41: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/41.jpg)
Techniques for Authentication
� Achieved by adding redundancy » authenticator, tag, etc., or» structure of message
� In some sense like Error Correcting Codes� Private Key - Public Key <=> Authentication -
Digital Signature� Attacks
» Substitution» Impersonation» Choice of above
![Page 42: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/42.jpg)
Digital Signature Authentication
Private keyHash
Digital SignatureOriginal
Forgery
FAIL!
Public key
![Page 43: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/43.jpg)
Authentication of Multimedia -New Issues
� Authentication of content instead of specific representation -» Example - JPEG or GIF image.
� Embedding of authenticator within content» Survive transcoding» Use existing formats
� Detect local changes» Simple block based authentication could lead to substitution
attacks
� Temporal relationship of multiple streams
![Page 44: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/44.jpg)
Fragile Watermarks
![Page 45: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/45.jpg)
Limitations of Fragile Watermarks
� Essentially same as conventional authentication – authenticate representation and not “content”.
� The differences being –» Embed authenticator in content instead of tag.» Treat data stream as an object to be “viewed” by
an human observer.» Computationally efficient?
![Page 46: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/46.jpg)
Feature Authentication
FeatureExtraction
Hash EncryptImage
Private Key
AuthenticatorFeatureSet
HashValue
Embed in perceptually irrelevant part of image
![Page 47: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/47.jpg)
Feature Authentication (contd.)
FeatureExtraction
Hash
Decrypt
Image
Public Key
FeatureSet
HashValue Same?
Authenticator Hash of FeatureSet of Original
Yes, Authentic
No, NotAuthentic
![Page 48: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/48.jpg)
Limitations of feature authentication
� Difficult to identify a set of definitive features.� Set of allowable changes has no meaningful
structure – certain “small changes” may not be allowed but the same time “large” changes may be allowed in other situations.
� “Strong” features facilitate forgeries.� “Weak” features cause too may false alarms.
![Page 49: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/49.jpg)
Difficulties with content Difficulties with content authentication of imagesauthentication of images
ŸContent is difficult to quantify.ŸMalicious (benign) modifications are difficult to quantify.
ŸImages considered as points in continuous space means there is not a sharp boundary between authentic and inauthentic images.
authentic inauthenticauthentic and inauthentic images which are similar to each other
![Page 50: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/50.jpg)
Distortion Bounded Authentication� Problem 1: allow flexibility in authentication to
tolerate small changes� Problem 2: to characterize and quantify the set
of allowable changes» Bound the errors» Perceptual distortion or pixel value distortion
� Provide “guarantees” against substitution attacks.
� Approach – bounded tolerance authentication» (semi-fragile)Watermarking techniques offer flexibility
but most do not offer bounds
![Page 51: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/51.jpg)
Distortion Bounded Authentication
� Quantize image blocks or features prior to computing authenticator.
� Quantization also done prior to verifying authenticity of image.
� Enables distortion guarantees – image considered authentic as long as change made does not cause quantized version to change.
� Can be used in many different ways
![Page 52: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/52.jpg)
Limitations
� Distortion added to “original” image. � Similar problems as feature authentication,
though to a lesser degree.� Significant changes may indeed be possible
within specified set of allowable changes.� How to define set of allowable changes?
![Page 53: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/53.jpg)
A Better Approach?A Better Approach?
original image
surely authentic images
fuzzy region
surely inauthentic images
Fuzzy region: authenticity of image is uncertain.
Chai Wah Wu - 2000
![Page 54: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/54.jpg)
Multimedia Fingerprinting
![Page 55: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/55.jpg)
Definitions
� A fingerprint is a characteristic of an object that can be used to distinguish it from other similar objects.» E.g., human fingerprints, marks on a fired bullet
� Fingerprinting is the process of adding fingerprints to an object or of identifying the fingerprint of an object that is intrinsic to an object.» Early examples: Table of logarithms with modified least
significant digits, maps drawn with slight deliberate variations. Thatcher documents.
� The advent of digital objects and their unauthorized distribution has lead to the need for novel fingerprinting techniques.
![Page 56: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/56.jpg)
Classification of Fingerprinting techniques (Wagner)
� Logical fingerprinting. » Object is digital. The fingerprints are computer-
generated and subject to computer processing.� Physical fingerprinting.
» This is the opposite of logical fingerprinting. Here the fingerprints depend on physical characteristics of the object.
![Page 57: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/57.jpg)
Classification of Fingerprinting techniques
� Perfect fingerprinting. » Any alteration to the object that will make the fingerprinting
unrecognizable must necessarily make the object unusable. Thus the distributor can always identify the recipient.
� Statistical fingerprinting. » Given sufficiently many misused objects to examine, the
distributor can gain any desired degree of confidence that he has correctly identified the compromised user. The identification is, however, never certain.
� Normal fingerprinting. » This is a catch-all category for fingerprinting that does not
belong to one of the first two categories.
![Page 58: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/58.jpg)
Classification of Fingerprinting techniques
� Recognition. » Recognize and record fingerprints that are already
a part of the object.� Deletion.
» The omission of some legitimate portion of the original object.
� Addition. » Legitimate addition» Modification.
![Page 59: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/59.jpg)
Classification of Fingerprinting techniques
� Discrete fingerprint. » An individual fingerprint with only a limited number
of possible values. – Binary fingerprint.– N-ary fingerprint.
� Continuous fingerprint. » Here a real quantity is involved and there is
essentially no limit to the number of possible values.
![Page 60: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/60.jpg)
Digital Fingerprints
� A mark is a position in an object that can be in one of a fixed number of different states (Boneh and Shaw)
» I.e., a codeword comprised of a number of letters from a preset alphabet
� A fingerprint is a collection of marks� Fingerprinting has two concerns
» How to mark an object» How to use these marks to create a fingerprint
� Fingerprinting cannot prevent unauthorized distribution, but acts as a deterrence mechanism by helping trace illegal copies back to source
» traitor: authorized users who redistribute content in an unauthorized manner
» traitor tracing: identifying traitors based on redistributed content
![Page 61: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/61.jpg)
Marking Assumption
The assumption states that a marking scheme designed to resist collusion and trace traitors with the following properties exist:
1. Colluding users may detect a specific mark only if the mark differs between their copies. Otherwise the mark cannot be detected.
If there is no collusion, fingerprint reduces to a serial number
2. Users cannot change the state of an undetected mark without rendering the object useless.
Basically, limits actions of colluding users
![Page 62: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/62.jpg)
Boneh-Shaw Construction
� Targeted at generic data with “Marking assumptions”(1998)» an abstraction of collusion model» E.g., assume a 6-bit content marked in the 2nd, 4th , and 5th
positions and let m1, m2 and m3 be the marked contents
» If m1, m2 and m3 collude the positions of the marks are determined» If m1and m2 collude only 4th and 5th marks can be identified
![Page 63: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/63.jpg)
Boneh-Shaw Construction
� Focus on tracing one of the colluders » Totally c-secure fingerprinting codes: Given a coalition of at most c
traitors, an illegal copy can be traced back to at least once traitor in the coalition.– Proved that for c>1 no such codes exist assuming colluder may leave
marks in unreadable state» Used randomization techniques to construct ε-error c-secure
codes that are able to capture at least one colluder, out of a coalition of c-colluders, with a probability of 1-ε for some small error rate of ε.
![Page 64: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/64.jpg)
Collusion Secure Codes
� Generate a code matrix whose rows are distinct fingerprints
� In the matrix, above the main diagonal is all ones and below is all zeros» May look like stairs, and the stairs width determine the ε, i.e.,
m1 : 111111111111m2 : 000111111111m3 : 000000111111m4 : 000000000111
» Prior to embedding each fingerprint is randomly permuted using afixed permutation
� A collusion will most likely generate a codeword different than m1, m2 , m3 and m4.
![Page 65: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/65.jpg)
Collusion Secure Codes (cont’d)
� Initially, fingerprints are far from each other (Hamming distance)
� The detector decodes the colluded fingerprint to nearestinitial fingerprint in the code matrix
� Arbitrarily small ε yields very long codes » Collusion resistance proportional to fourth root of content size (i.e.,
to capture at least one of c-colluders code length must be of the order O(c4logc)).
� Lot of follow up work in crypto literature that extends and improves Boneh-Shaw results.
![Page 66: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/66.jpg)
Embedded Fingerprinting for Multimedia
embedembedDigital Fingerprint
Multimedia Document
101101 …101101 …
Customer’s ID: Alice
Distribute to Alice
Fingerprinted CopyFingerprinted Copy
embedembedDigital Fingerprint
Multimedia Document
101101 …101101 …
Customer’s ID: Alice
Distribute to Alice
Fingerprinted CopyFingerprinted Copy
Collusion Attack Collusion Attack (to remove fingerprints)(to remove fingerprints)
AliceAlice
BobBob
Colluded CopyColluded Copy
Unauthorized Unauthorized rere--distributiondistribution
�Fingerprinted docfor different users
Collusion Attack Collusion Attack (to remove fingerprints)(to remove fingerprints)
AliceAlice
BobBob
Colluded CopyColluded Copy
Unauthorized Unauthorized rere--distributiondistribution
�Fingerprinted docfor different users
Extract Extract FingerprintsFingerprints
Suspicious Suspicious CopyCopy
101110 …101110 …
Codebook
Alice, Bob, …
Identify Identify TraitorsTraitors
Extract Extract FingerprintsFingerprints
Suspicious Suspicious CopyCopy
101110 …101110 …
Codebook
Alice, Bob, …
Identify Identify TraitorsTraitors
�������������������������� �� �� ��
��������������������
���� ���� ���� ����
���������������������������� ���� ���� ����
����������������������������
������� ������� ������� ������� ��������������������
![Page 67: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/67.jpg)
What is Different?
� New issues with multimedia» “Marking assumptions” do not directly carry over …» Some code bits may become erroneously decoded due to strong
noise and/or inappropriate embedding» Can choose appropriate embedding to prevent colluders from
arbitrarily changing the embedded fingerprint bits � Want to trace as many colluders as possible� Major Concerns
» How to embed/detect the fingerprint– Deploy techniques from watermarking
» How to generate the fingerprint– Utilize techniques from coding theory
» The type of attack the fingerprinted object undergoes
![Page 68: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/68.jpg)
Marking Assumption for Multimedia Fingerprinting
Marking assumption considers a scheme with two specific requirements� Fidelity requirement (Easy to satisfy)
» Marks are perceptually invisible and can be discovered only by comparison
» Unmarked object is not available� Robustness requirement (Difficult to achieve)
» Undetected marks cannot be altered or removed
![Page 69: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/69.jpg)
Spread-Spectrum Fingerprint Embedding/Detection
� Spread-spectrum embedding/detection» Provide very good tradeoff on imperceptibility and
robustness, esp. under non-blind detection» Typical watermarking-to-noise (WNR) ratio: -20dB in
blind detection, 0dB in non-blind detection� Embedding: X=S+αW where S is the original
object, Wi is the fingerprint, and α is the embedding strength
� Detection: Analysis of the similarity between Yand Wi, i.e., correlation(Y,Wi) or correlation(Y-S,Wi)
![Page 70: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/70.jpg)
Fingerprinting Generation
� Choice of modulation schemes
jj uw =
∑=
=B
iiijj b
1uw
{ }1,0∈ijb { }1±∈ijb
# of fingerprints = # of orthogonal bases
# of fingerprints >> # of orthogonal bases
Orthogonal modulation
(Binary) coded modulationfor or
1st bit1st bit 2nd bit2nd bit
......
![Page 71: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/71.jpg)
Attacks on Fingerprinting Systems
� Attacks on the marking system» Exploiting the robustness of the fingerprinting embedding and
detection scheme» Collusion attacks. Collusion: Y=g(X1,X2, …Xk) where g(.) is a
function designating the nature of modification on collection offingerprinted objects available.
– More effective– May yield even better quality than the distributed object
� The traitor may have two types of goal1. Removal of the fingerprints from the fingerprinted-object2. Framing an innocent user
� Design Goal: Improve collusion resistance w.r.t. type-1 while increasing robustness to type-2 attacks.
![Page 72: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/72.jpg)
Collusion Attacks by Multiple Users
� Interesting collusion attacks become possible� Fairness: Each colluder contributes equal share through
averaging, interleaving, and nonlinear combining
Colluded copy
Collusion byaveraging
Originally fingerprintedcopies
1/3
Alice ChrisBob
Colluded copy
Originally fingerprintedcopies
Alice Chris
Cut-and-paste attack
![Page 73: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/73.jpg)
Linear vs. Nonlinear Collusion
� Linear collusion by averaging is simple and effective� Colluders can output any value between the minimum and
maximum values, and have high confidence that such spurious value is within the range of JND. � Important to consider nonlinear collusion as well.
� Order statistics based nonlinear collusions
( )m in m a x
m in m a x m in m a x
m o d m in m a x
m in
m a x
; ; ;
,
w .p . w .p . 1
a v e m e d ia nj j j j
j j j
n e g m e dj j j j
jr a n d n e gj
j
V V V V
V a v e r a g e V V
V V V V
V pV
V p
=
= + −
= −
CC Skk
jjjSkk
jj wgJNDxygV ∈∈ ⋅+== )()( )()(
![Page 74: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/74.jpg)
(Image) Steganographyand Steganalysis
![Page 75: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/75.jpg)
Steganography
� Steganography - “covered writing”.� For example (sent by a German spy during
World War I),
Apparently neutral's protest is thoroughly discounted and ignored. Isman hard hit. Blockade issue affects pretext for embargo on byproducts, ejecting suetsand vegetable oils.
Pershing sails from NY June I.
![Page 76: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/76.jpg)
Ancient SteganographyHerodotus (485 – 525 BC) is the first Greek historian. His great work, The Histories, is the story of the war between the huge Persian empire and the much smaller Greek city-states.
Herodotus recounts the story of Histaiaeus, who wanted to encourage Aristagoras of Miletus to revolt against the Persian king. In order to securely convey his plan, Histaiaeus shaved the head of his messenger, wrote the message on his scalp, and then waited for the hair to regrow. The messenger, apparently carrying nothing contentious, could travel freely. Arriving at his destination, he shaved his head and pointed it at the recipient.
![Page 77: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/77.jpg)
Ancient SteganographyPliny the Elder explained how the milk of the thithymallus plant dried to transparency when applied to paper but darkened to brown when subsequently heated, thus recording one of the earliest recipes for invisible ink.
Pliny the Elder. AD 23 - 79
The Ancient Chinese wrote notes on small pieces of silk that they then wadded into little balls and coated in wax, to be swallowed by a messenger and retrieved at the messenger's gastrointestinal convenience.
![Page 78: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/78.jpg)
Renaissance Steganography
Giovanni Battista Porta(1535-1615 )
Giovanni Battista Porta described how to conceal a message within a hard-boiled egg by writing on the shell with a special ink made with an ounce of alum and a pint of vinegar. The solution penetrates the porous shell, leaving no visible trace, but the message is stained on the surface of the hardened egg albumen, so it can be read when the shell is removed.
![Page 79: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/79.jpg)
Modern Steganography - The Prisoners’ Problem
� Simmons – 1983� Done in the context of USA – USSR nuclear
non-proliferation treaty compliance checking.
Wendy
Hello Hello
“Hello”
![Page 80: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/80.jpg)
Yes
Modern Terminology and (Simplified) Framework
NoEmbedding Algorithm
CoverMessage
Stego Message
SecretKey
SecretMessage
Message Retrieval Algorithm
Secret Message
Secret Key
Is Stego Message?
Suppress Message
Alice Wendy Bob
![Page 81: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/81.jpg)
Secret Key Based Steganography
� If system depends on secrecy of algorithm and there is no key involved – pure steganography» Not desirable. Kerkhoff’s principle.
� Secret Key based steganography� Public/Private Key pair based steganography
![Page 82: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/82.jpg)
Active and Passive Warden Steganography
� Wendy can be passive:» Examines all messages between Alice and Bob. » Does not change any message» For Alice and Bob to communicate, Stego-object should be
indistinguishable from cover-object.
� Wendy can be active:» Deliberately modifies messages by a little to thwart any
hidden communication.» Steganography against active warden is difficult.» Robust media watermarks provide a potential way for
steganography in presence of active warden.
![Page 83: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/83.jpg)
Steganalysis
� Steganalysis refers to the art and science of discrimination between stego-objects and cover-objects.
� Steganalysis needs to be done without any knowledge of secret key used for embedding and maybe even the embedding algorithm.
� However, message does not have to be gleaned. Just its presence detected.
![Page 84: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/84.jpg)
Cover Media
� Many options in modern communication system:» Text» Slack space» Alternative Data Streams» TCP/IP headers» Etc.
� Perhaps most attractive are multimedia objects -– Images– Audio– Video
� We focus on Images as cover media. Though most ideas apply to video and audio as well.
![Page 85: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/85.jpg)
Steganography, Data Hiding and Watermarking
� Steganography is a special case of data hiding.» Data hiding in general need not be
steganography. Example – Media Bridge.� It is not the same as watermarking.
» Watermarking has a malicious adversary who may try to remove, invalidate, forge watermark.
� In Steganography, main goal is to escape detection from Wendy.
![Page 86: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/86.jpg)
Information Theoretic Framework
� Cachin defines a Steganographic algorithm to be secure if the relative entropy between the cover object and the stego object pdf’s is at most :
� Perfectly secure if � Example of a perfectly secure techniques known but not
practical.
ε
ε
)
0=ε
![Page 87: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/87.jpg)
Steganography in Practice
Image Noise
Content
ModulatedMessage
SecretMessage
Stego Image+
![Page 88: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/88.jpg)
Steganalysis in Practice
� Techniques designed for a specific steganography algorithm» Good detection accuracy for the specific technique» Useless for a new technique
� Universal Steganalysis techniques» Less accurate in detection» Usable on new embedding techniques
![Page 89: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/89.jpg)
Simple LSB Embedding in Raw Images
� LSB embedding» Least significant bit plane is changed. Assumes
passive warden.� Examples: Encyptic, Stegotif, Hide� Different approaches
» Change LSB of pixels in a random walk» Change LSB of subsets of pixels (i.e. around
edges)» Increment/decrement the pixel value instead of
flipping the LSB
![Page 90: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/90.jpg)
LSB Embedding
![Page 91: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/91.jpg)
Steganalysis of LSB Embedding
� PoV steganalysis - Westfeld and Pfitzmann.» Exploits fact that odd and even pairs from “closed
set” under LSB flipping. » Accurately detects when message length is
comparable to size of bit plane.
� RS-Steganalysis - Fridrich et. al. [14]
» Very effective. Even detects around 2 to 4% of randomly flipped bits.
![Page 92: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/92.jpg)
LSB steganalysis with Primary Sets
� Proposed by Dumitrescu, Wu, Memon » Based on statistics of sets defined on neighboring pixel
pairs.» Some of these sets have equal expected cardinalities, if
the pixel pairs are drawn from a continuous-tone image.» Random LSB flipping causes transitions between the sets
with given probabilities, and alters the statistical relations between their cardinalities.
» Analysis leads to a quadratic equation to estimate the embedded message length with high precision.
![Page 93: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/93.jpg)
State Transition Diagram for LSB Flipping
X(2k-m,2k)
(2k+1+m,2k+1)
W (2k+1,2k) (2k,2k+1)
Z (2k,2k)
(2k+1,2k+1)
V (2k+1+m,2k) (2k-m,2k+1)
10,01
11,01
11,0110
,01
00,1
0
00,1100,1100
,10
m≥1,k≥0Y
(2k+m,2k) (2k+1-m,2k+1)
X,V, W, and Z, which are called primary sets
![Page 94: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/94.jpg)
Transition Probabilities
� If the message bits of LSB steganography are randomly scattered in the image, then
� Let X, Y, V, W and Z denotes sets in original image and X’, Y’. W’ and Z’ denote the same in stego image.
( )
( ) ( )
( ) .2
11
,2
12
1001
,2
100
2
2
=
−==
−=
p
pp
p
ρ
ρρ
ρ
iii)
ii)
i)
![Page 95: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/95.jpg)
Message Length in Terms of Cardinalities of Primary Sets
� Cardinalities of primary sets in stego image can be computed in terms of the original
� Assuming
� Where
221'
221'
pXpVV
pVpXX
+
−=
+
−=
−+
+−=2
12
1'2 ppZppWW
and some algebra, we get: }{}{ YEXE =
( ) 0'''25.0 2 =−+−+⋅ XYpPXpγ
.'' ZW ∪=γ .ZW ∪=
![Page 96: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/96.jpg)
Simulation Results
![Page 97: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/97.jpg)
Embedding in JPEG Images
� Embedding is done by altering the DCT coefficient in transform domain
� Examples: Jsteg, F5, Outguess� Many different techniques for altering the
DCT coefficients
![Page 98: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/98.jpg)
F5
� F5 uses hash based embedding to minimize changes made for a given message length
� The modifications done, alter the histogram of DCT coefficients
� Given the original histogram, one is able to estimate the message length accurately
� The original histogram is estimated by cropping the jpeg image by 4 columns and then recompressing it
� The histogram of the recompressed image estimated the original histogram
![Page 99: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/99.jpg)
F5 plot
Fig. 5. The effect of F5 embedding on the histogram of the DCT coefficient (2,1).
![Page 100: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/100.jpg)
Outguess
� Embeds messages by changing the LSB of DCT coefficients on a random walk
� Only half of the coefficients are used at first� The remaining coefficients are adjusted so
that the histogram of DCT coefficient would remain unchanged
� Since the Histogram is not altered the steganalysis technique proposed for F5 will be useless
![Page 101: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/101.jpg)
Outguess
� Researchers proposed “blockiness” attack � Noise is introduced in DCT coefficients after
embedding� Spatial discontinuities along 8x8 jpeg blocks
is increases� Embedding a second time does not introduce
as much noise, since there are cancellations� Increase or lack of increase indicates if the
image is clean or stego
![Page 102: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/102.jpg)
Universal Steganalysis Techniques
� Techniques which are independent of the embedding technique
� One approach – identify certain image features that reflect hidden message presence.
� Two problems» Calculate features which are sensitive to the embedding
process» Finding strong classification algorithms which are able to
classify the images using the calculated features
![Page 103: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/103.jpg)
What makes a Feature “good”
� A good feature should be:» Accurate
– Detect stego images with high accuracy and low error
» Consistent – The accuracy results should be consistent for a set of
large images, i.e. features should be independent of image type or texture
» Monotonic– Features should be monotonic in their relationship with
respect to the message size
![Page 104: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/104.jpg)
IQM
� IQM’s can be used as features � From a set of 26 IQM measures a subset with
most discriminative power was chosen� ANOVA is used to select those metrics that
respond best to image distortions due to embedding
![Page 105: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/105.jpg)
IQM
Scatter plot of 3 image quality measures showing separation of marked and unmarked images.
![Page 106: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/106.jpg)
Classifiers
� Different types of classifier used by different authors. » MMSE linear predictor » Fisher linear discriminates as well as a SVM
classifier� SVM classifiers seem to do much better in
classification� All the authors show good results in their
experiments, but direct comparison is hard since the setups are very much different.
![Page 107: An Introduction to Basic Concepts and Watermarking and ...yao/EE4414/memon_F05_v2.pdfAuthentication of Multimedia - New Issues Authentication of content instead of specific representation](https://reader035.vdocument.in/reader035/viewer/2022062509/610a3f70ddc14b05e93427cb/html5/thumbnails/107.jpg)
So What Can Alice (Bob) Do?
� Limit message length so that detector does not trigger
� Use model based embedding.» Stochastic Modulation
� Adaptive embedding » Embed in locations where it is hard to detect.
� Active embedding» Add noise after embedding to mask presence.» Outguess