VIDEO WATERMARKING

Presented by:-
Subhajit Brojabasi

Guided by:-
prof. Mihir singh
1
CONTENTS

• Introduction
• Need of video watermarking
• Video watermarking terminologies
• Types
• Desired properties
• Application of watermarking
Structure of a Typical Video Watermarking
System
• An algorithm of video watermarking
• Attack on watermark
• Applications
• Conclusion
• References
2
INTRODUCTION

Definition: A digital watermark is a digital signal or

pattern inserted into a digital document such as
text, graphics or multimedia, and carries
information unique to the copyright owner.

 Video watermarking is an extension of this

concept.

3
Need of video watermarking
 Protect copyright of a data.
 Video Watermarking can help
◦ Prove ownership
◦ Identify a misappropriating person
◦ Trace the video dissemination

 Video watermarking introduces some issues not

present in image watermarking
 Frame averaging, frame dropping, frame
swapping …

4
Video watermarking terminologies

• Digital Video

• Payload

• Perceptibility

• Robustness

• Security

5
TYPES

6
 DESIRED PROPERTIES

 Robust
Video watermarking ( frame averaging, frame
dropping, frame swapping, statistical analysis,
interpolation... )
 Recoverable from a document
 Removable by authorized users
 Low error probability
 Provide the original information embedded
reliably
 Capacity
 Invisible

7
 APPLICATION OF WATERMARKING

• Owner Identification
• Proof of Ownership
• Transaction Tracking
• Content Authentication
• Broadcast Monitoring
• Copy Control
• Device Control
• Automatic monitoring of copyrighted material on
the Web

8
A Generic approach to watermarking digital
data

Thereare 3 main processes involved in

watermarking —

1. Insertion of a watermark

2. Detection of a watermark

3. Removal of a watermark

9
Inserting a watermark

 It consists of :-

A watermark insertion unit that uses:-

a. Original video
b. Watermark
c. A user key

to obtain the watermarked video.

10
watermark insertion unit
 The user key , input video and the watermark
are passed through a watermark insertion unit.

 This results in a watermarked video.

11
Watermark Extraction Unit
 Two phases:-

1. Locating the watermark

2. Recovering the watermark information.

12
Watermark Detection Unit
Consists of :-

a. An extraction unit to first extract the

watermark.

b. Compare it with the original watermark

inserted.

c. The output is Yes or No depending on

whether the watermark is present.

13
14
Current Techniques — The LSB technique
 Let’s consider still images.

 Inthis technique least significant bit of each pixel

is used to embed the watermark or the copyright
information.

 Each
pixel of the color image has three
components — red, green and blue.

3 bytes for each pixel.

 Each color has 1 byte, or 8 bits.

15
16
The LSB technique…….
A pixel that is bright purple in color can
be shown as

X0 = {R=255, G=0, B=255}

Another pixel:

X1 = {R=255, G=0, B=254}

17
The algorithm
Let W be watermarking information
For every pixel in the image, Xi
Do Loop:
Store the next bit from W in the LSB
position of Xi [red] byte
Store the next bit from W in the LSB
position of Xi [green] byte
Store the next bit from W in the LSB
position of Xi [blue] byte
End Loop

18
Distortions and attacks
Additive Noise.

Cropping.

Compression.

Statistical Averaging.

Multiple Watermarking.

Mosaic Attack
19
20
Current VIDEO WATERMARKING Techniques

◦ (DWT) Discrete Wavelength Transform

◦ (PCA) Principle Component Analysis

• (QIM) Quantization Index Modulation

Watermarking I-frame (Mpeg-1,2)

Video object watermarking (Mpeg-4)

21
Discrete Wavelet Transform
Normalized to 256 X 256 pixel size

Perform DWT for each frame

22
Advantage of DWT
 Prevent from frame dropping
 Prevent from adding noise / another video

Mark with m1 Mark with m3 Mark with m0 Mark with m6

Scene change occur

23
Watermark Preprocess
 Safe from frame averaging
 Hard to reconstruct the watermark

Place the bit-

planes side by
side
85 Bit
17
m0
64 Decomposition 0
0 4 0 4
Crop 1 5 25 1 5
2 6 8 bit-
6 2 6
3 7 planes 3 7

17
m7
0
0 4
25 1 5
8 bit-
6 2 6
planes 3 7
Total 8 images is
obtained

24
Watermark Embedding

C[i] = C[i] + 1 , if W[j] = 1

C[i] = C[i] , otherwise
LL, HH is not watermark

LL HL 170
LH HH
HL3
HL2 0 4
LH3 HH3
HL1
256 1 5
LH2 HH2 2 6
3 7
Watermark mi

LH1 HH1

25
Watermark Detection

Mark with m1 Mark with m3 Mark with m0 Mark with m6

Scene change occur

1. Video Frame
3. DWT
256 X 256 2. Detect the scene change

4. Compare
Original video frames

DWT result frames

5. Averaging the watermark
get from different frames

170
85 6. Combine 8 bit-planes 0 4
64 Watermark mi 256 1 5
7. Get different part of W 2 6
from different scene and 3 7
m0 m4 reconstruct the large image Extracted Original
256 m1 m5
m2 m6
8. Correlate and verify E O
m3 m7
170

26
ATTACK ON WATERMARK

Processing operation • Frame averaging

Compression • Frame dropping
Rescaling • Frame swapping
Cropping • Statistical analysis
Multiple watermarks • Interpolation

27
Future Scope
 Implement and use do experiment on the
scheme

 Identifiable from very short segment of video

 Change to blind watermarking scheme

 Make use of the object

28
 CONCLUSION
Need due to Copyright protection
Different from image watermarking
Brief review of current technologies
DWT-based Video Watermarking
Scheme with Scramble Watermark
Future Direction
Inadequate for general use
Different tools offering this
Set to grow
29
REFERENCES

 F. Hartung and M. Kutter, “Multimedia watermarking

techniques”,Proceedings of the IEEE, vol. 87, no. 7 , July1999.

 I.Cox, J. Kilian, T. Leighton, and T. Shamoon, “Secure spread

spectrum watermarking for images, audio and video,” in Proc.
IEEE Int. Conf. Image Processing (ICIP 96), Lausanne,
Switzerland, Sept. 1996.

 Yeo and M.M. Yeung, “Analysis and synthesis for new digital
video applications,” icip, International Conference on Image
Processing (ICIP'97),vol. 1, pp.1,1997.

30
31