Data Hiding in Image and Video: Part II

1
Data Hiding in Image and VideoPart IIDesigns
and Applications

• Min Wu, Heather Yu, and Bede Liu

2
Outlines

• Introduction
• Multilevel Data Hiding in Grayscale Image
• Multilevel Data Hiding in Video
• Conclusion

3
Introduction

• Goal
• apply the solutions in Part I to specific design
  problems and present details of embedding data

4
Multilevel Data Hiding in Grayscale Image

• Introduction
• Spectrum Partition
• System Design
• Experimental Results

5
Multilevel Data Hiding in Grayscale Image --
Introduction

• Present a two-level data hiding using two types
  of embedding mechanisms
• Basis Fig5. in Part I
• Basic Assumptions/Conditions
• Grayscale Images
• Embedding Domain 88 block DCT coefficients
• Using Spectrum Segments for Embedding
• Dealing with non-coherent case

6
Multilevel Data Hiding in Grayscale Image --
Introduction
7
Spectrum Partition

• Data Model and Formula
• Experimental Results

8
Spectrum Partition-Data Model(1)

• Embedding
• where
• the watermark s1, , sn is an n-sample known
  sequence,
• b a bit to be embedded and is equally likely to
  be -1 or 1,
• di noise, i.i.d. Gaussian

9
Spectrum Partition-Data Model(2)

• A few considerations
• Bits can be embedded in all bands. In many cases,
  bits are embedded in mid-band due to
• Low band coefficients generally have higher power
• High band coefficients are vulnerable to attacks
• Noise Model can be extended to Normal
  Distribution with Various Covariance. Whitening
  should be performed in such cases

10
Spectrum Partition-Data Model(3)

• The detector
• The mean

11
Spectrum Partition-Simulation(1)

• Subject 141 Images
• Embedding the Block-DCT spread spectrum
  algorithm proposed by Podilchuk-Zeng
• Detection the q-statistic proposed by Zeng-Liu
• Three watermarks are used
• Pre-processing
• An estimation of the host signals power is
  performed based on testing images
• A set of known signals are added to help locating
  host signal from noise

12
Spectrum Partition-Simulation(2)

• Detection Defined two statistics q and q,
  with and without the weighting

13
Spectrum Partition-Simulation(3)

• Experiments
• DCT coefficients are ordered in zig-zag order
• Several distortion are introduced while computing
  q-statistics
• JPEG with different quality factors
• Low pass filtering
• q-statistics are normalized with respect to
  number of embeddable coefficients, see Figures
• Q is maximum when the embedding starts around
  6-11
• Q is larger than q and its monotone
• Conclusion
• For high robustness, embed the bit to mid-band
  coefficients
• For high payload, embed the bit to low-band
  coefficients

14
Spectrum Partition-Simulation(4)
15
Spectrum Partition-Simulation(5)
16
Spectrum Partition-Simulation(6)
17
System Design

• Block Diagram
• Two Level Embedding

18
System Design Block Diagram(1)

• Embedding

19
System Design Block Diagram(2)

• Detecting

20
Two Level Embedding(1)

• First Level
• Using Odd-Even Embedding in the Low Band
• Quantization Techniques are applied

21
Two Level Embedding(2)

• Second Level
• Using Type I Spread Spectrum Technique
• Antipodal Modulation Is Used
• where vi original coefficients
• vi marked coefficients
• b antipodal mapping from b, which is 1 or
  1
• watermark strength, adjusted by the
  just-noticeable- difference (JND) standard

22
Experimental Results
23
Multilevel Data Hiding in Video

• Embedding Domain
• Variable Embedding Rate (VER) Versus Constant
  Embedding Rate (CER)
• Control Data Versus User Data
• Experimental Results

24
Embedding Domain(1)

• Problems Introduced by Consecutive Frames
• Add/Drop Some Frames
• Switch the Order of Frames
• Generate New Frames
• Possible Attacks
• Collusion Attack
• Solution
• Adding Redundancy

25
Embedding Domain(2)

• To Avoid Frame-Jitter
• Partitioning the Video into Temporal Segments
• Embedding Same Data in Every Frame of a Segment

26
Embedding Domain(3)

• To Avoid Frame Drop, Reordering, Insertion
• Embedding the Same User Data As Well As a Shorten
  Version of Segment Index
• The Segment Index Is Part Of the Control Bits

27
Variable Embedding Rate (VER) vs. Constant
Embedding Rate (CER)

• Problem
• The Uneven Embedding Capacity Arises Both From
  Region to Region within a Frame and From Frame to
  Frame
• Solution
• Combine VER and CER
• The Intra-Frame Unevenness Is Handled by CER and
  Shuffling
• The Inter-Frame Unevenness Is Handled by VER and
  Additional Side Information

28
Number of Bits Embedded in Each Frame

• Number of Bits That Can Be Embedded in Each Frame
  Changes Greatly
• Estimate Number of Bits for Each Frame
• Estimate the Achievable Embedding Payload
• Based on Energy of DCT Coefficients, Number of
  Embeddable Coefficients
• Set Two Threshold and
• If do not embed data
• If a number
  of bits are embedded
• If bits are embedded in
  higher rate

29
Estimation of Payload

• For Type I Spread Spectrum Embedding,
• The Mean of Detection Statistic Is
• Bit Error Probability Is Given by
• Maximum Bit Error Probability Is Given by
• A Lower Bound of Mean Detection Statistic Is
  Defined by
• The Detection Statistic When All Embeddable
  Coefficients Are Used Is Given By
• The Payload Is

30
Control Data Versus User Data(1)

• Control Data Additional Information
• Include Frame Sync Index, Number of Bits Embedded
  in Each Frame
• Embedding Frame Sync
• A Short Version of Video Segment Index
• Assume Frame Syncs Range is 0 to K-1
• The i-th Segment Is Labeled as

31
Control Data Versus User Data(2)

• User Data Information
• TDM with Shuffling IS Applied
• Orthogonal Modulation Is Used to Double the
  Number of Embedded Bits
• Assume 2B bits Are Embedded

32
Block Diagram
33
Experimental Results
34
Conclusion

• Demonstrate How to Apply General Solutions in
  Part I to Specific Designs
• Made use of
• Two types of Embedding
• Modulation and Multiplex Techniques
• Shuffling
• Multilevel Data Hiding