Video Watermarking Algorithm based on Combined Transformation
and Digital Holography
De Li
1
, XueZhe Jin
1
and Jongweon Kim
2
1
Department of Computer Science, Yanbian University, 133002, Yanji, China
2
Department of Contents and Copyright, Sangmyung University, 03016, Seoul, Korea
Keywords: Digital Hologram, 2D DWT, 3D DCT, DFRNT.
Abstract: Recent researches of robust digital watermarking are focused on the processing method in frequency field.
Most of those algorithms are limited on one transform domain or combine different kinds of transform field
together. In this paper, a new video watermarking algorithm is proposed, which is applied in 2D DWT and
3D DCT combined with Discrete Fraction Random Transform (DFRNT) encryption and digital holography
respectively. The embedded information is applied in 2D bar-code which obtained from holograph.
Extraction of the watermark uses the tech of holographic reconstruct. The results of experiment show that
the proposed watermarking algorithm has a good performance of robustness and security.
1 INTRODUCTION
With the rapid development of Internet, more and
more people share information and it is easier to
store, copy and transmit the multi-media data than
before. The evolution of network brings the great
advancement of the information society but it also
causes some challenges such as copyright protection.
More and more researchers pay high attention to
solve the digital copyright issues and it is not
acceptable to distort, transmit and invade
copyrighted material. The anti-attack watermarking
algorithm for video is proposed in this paper. It
shows good performance of concealing，anti-attack
and security.
Newly developed digital watermarking
algorithms are, wavelet transform with Singular
Value Decomposition (SVD) (Bao, 2005), spread
spectrum based on HVS (Zolghadrashli, 2007),
Pseudo-3-D DCT and Quantization (Huang, 2010),
1D Discrete Fourier Transform (DFT) and Radon
Transform (Liu, 2010), compression domain method
(Ling, 2010. Gujjunoori, 2013)
and adaptive method
like Particle Swarm Optimization (PSO) (Margarita,
2009. Wang, 2010). Most recent watermarking
algorithms applying binary or gray image for
watermarking have limitation in the quantity of
embedded information. Besides, various attacking
methods have been developed constantly. Based on
these researches, it was proposed that combine the
optical technology such as visual optic with digital
holography into the watermarking algorithm and got
some positive achievement (Liang, 2006). But it was
still not a perfect method which can resist all kinds
of attack with strong performance of security. So, a
video watermarking algorithm with robustness and
security is hot issue. This paper proposes an
algorithm that combine the tech of combined
frequency domain processing, random exchange of
discrete fraction and digital holography together.
2 CORRELATION THEORIES
2.1 Digital Holography
Digital holography is a tech that applying CCD to
exchange board to record holography image and
emulating diffraction to reconstruct the image. The
whole reconstructing process contains record, store,
process and reproduce. Recent researches focus on
optical technology combining with information
security. Especially, the field of digital
watermarking has great potential. In 2002, Takai and
his team (Takai, 2002) proposed an algorithm of
digital holography getting the watermarking image,
which is measured by holography tech and embed
the watermark immediately in space domain and
Li, D., Jin, X. and Kim, J.
Video Watermarking Algorithm based on Combined Transformation and Digital Holography.
DOI: 10.5220/0005749803790383
In Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2016) - Volume 3: VISAPP, pages 381-385
ISBN: 978-989-758-175-5
Copyright
c
2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
381
recover the original watermarking image. But the
robustness of their algorithm will deteriorate if the
image is under a low-pass filtering attack. Except
low-pass filtering attack, the digital holography has
good characteristics of encryption, irrevocability and
cropping.
Figure 1 is the process of computer-generated
hologram.
Figure 1: Process of computer-generated hologram.
Figure 2 shows generation and reconstruction of
the hologram.
(a) Original image (b) Holographic image
(c) Holographic representation (d) Cutting holographic image
(e) Reproduced image (f) Extracted image
Figure 2: Generation and reconstruction of the hologram.
2.2 Random Exchange of Discrete
Fraction
The method of Random Exchange of Discrete
Fraction Transform(REDFT) is based on DFT and it
has better mathematical performance than DFT. It
can be satisfied in linearity, unitarity, additivity,
cyclicity and self-randomicity (Zhang, 2012).
The 1D and 2D signal of REDFT can be
expressed as below.
(1)
(2)
is the kernel matrix of REDFT.
V
(3)
α
is factorial of discrete fraction,
T
is the cycle
time of
α
,
is a diagonal matrix which represent
a group of eigenvalues exp
πnα
:n
0,1,2,…,N 1.
can be expressed as formula(4).
1,exp
,…,exp
(4)
The matrix
V
is formed from the eigenvector
,
,…,
and the eigenvector can be get from
Schmidt standard normalization.
Matrix
V
can be expressed as equation (5).
V
,
,…,
(5)
is eigenvector of diagonal random matrix Q
and Matrix Q can be get from NN random matrix
P.
QP
/2
(6)
Matrix V satisfied below relation.
V
(7)
The result of the transform will differentiate by
the matrix P and the eigenvector is random vector.
The random matrix contains N(N+1)/2 isolated
elements and the security can be expressed as
2N(N+1)/2. It means the method itself is an
encryption procedure and the key is very difficult to
decipher. Even though the adversary knows the
transform procedure, there is little possibility of
obtaining the embedded information.
3 WATERMARKING
ALGORITHM AND RESULTS
ANALYSIS
3.1 Watermarking Algorithm
If the watermark is embedded in high frequency
field less sensitive for visual, it will be damaged
during the attack of transmission. On the contrary it
is embedded in low frequency field, the quality of
the image will deteriorate. The proposed algorithm
will choose an intermediate field to embed the
watermark information. The algorithm uses 2D
wavelet transform and 3D DCT first and performs
DFRNT scramble to get the watermark information.
The robustness and security of the algorithm has
VISAPP 2016 - International Conference on Computer Vision Theory and Applications
382
improved by the technology of off-axis holography.
The process of the algorithm is shown in figure
3.
Figure 3: Process of the algorithm.
3.2 Experimental Results and Analysis
The original video for this experiment is 100 frames
which are uncompressed and the image size for
watermarking is 21×21 2D barcode
image. Bit
Error Rate (BER) performance evaluations are
measured applying various attacks.
(1) Gaussian noise attacks
Table 1: Experimental results of Gaussian noise attacks.
Variance Result Extraction BER
0.001
0.00
0.005
0.01
0.01
0.05
Table 1 shows that the video frame is not enough
clear after Gaussian noise attack but we can extract
the watermarking. It means that the algorithm can
resist the Gaussian noise attack.
(2) Salt & pepper noise attacks
Table 2: Experimental results of salt & pepper noise
attacks.
Density Result Extraction BER
0.005
0.00
0.01
0.02
0.05
0.06
Table 2 shows that the extraction of
watermarking still works enough after the salt &
pepper attack. It means the algorithm has a good
performance to resist salt & pepper attack.
(3) Low-pass filtering attack
Low-pass filtering attack impairs the edge of the
video frame so that the image is blurred.
Table 3: Experimental results of low-pass filtering attacks.
Frequency Result Extraction BER
55
0.04
75
0.02
95
0.00
Video Watermarking Algorithm based on Combined Transformation and Digital Holography
383
Table 3 shows that the filtering attack has little
effect on the quality of the video frame.
(4) Rotation attack
During the transmission, the video file can be
attacked by rotation. In this paper, the angle of
rotation attack is from 1 to 90 degrees.
Table 4: Experimental results of rotation attacks.
Rotation
angle
Result Extraction BER
10°
0.05
45°
0.13
60°
0.10
Table 4 shows that the extraction of the
watermark can be performed successfully in three
cases of rotation angle: 10, 45 or 60 degrees.
(5) Frame attack
Frame attack is exchange the order of frame, add or
delete frame so that the user can not be able to get
the correct video file. Test results of robustness to
frame attack is shown in table 5.
Table 5: Experimental results of frame attacks.
Frame exchange Extraction BER
1
st
↔ 9
th
0.00
5
th
↔ 21
st
0.01
29
th
↔ 53
rd
0.02
Table 5: Experimental results of frame attacks (cont.).
Lost frames Extraction BER
3
0.00
5
0.01
15
0.02
(6) Compression attack
Most video files on Internet are compressed by
MPEG. The compressed video file in MPEG has a
good performance but the loss of information cannot
be avoid 100%. MPEG-4 compressed video files are
tested and the performance is evaluated. Table 6
shows the result of compression attack by different
bit rate. Even if the bit rate goes up to 500Kbps, the
extraction still can be clear to clarify.
Table 6: Experimental results of compression attacks.
Standard MPEG-4
Bit rate 256Kbps 512Kbps
Extraction
BER 0.19 0.07
The experimental results show that the algorithm
combining DFRNT with digital holography has a
good performance to endure all kinds of attack and
the encryption feature of DFRNT provide a good
security than other algorithm.
4 CONCLUSIONS
The proposed algorithm shows good performance of
robustness and security. The algorithm combines 2D
wavelet transform and 3D DCT and then apply
DFRNT for scramble. During the construction of
watermarking the 2D bar-code information will be
added to the video file as the mode of holography
image. The extraction of watermarking takes
holograph-reproduce which can improve security of
the algorithm. The result of experiment shows good
performance to various attacks and reasonable
improvement of robustness and security.
VISAPP 2016 - International Conference on Computer Vision Theory and Applications
384
ACKNOWLEDGEMENTS
This research is supported by Ministry of Culture,
Sports and Tourism(MCST) and Korea Creative
Content Agency(KOCCA) in the Culture
Technology(CT) Research & Development Program
2015.
REFERENCES
P, Bao. and X, Ma. 2005. Image Adaptive Watermarking
Using Wavelet Domain Singular Value
Decomposition. IEEE Transactions on Circuits and
Systems for Video Technology, 15:96~102.
A, Zolghadrashli. and S. Rezazadesh, 2007. Evaluation of
Spread Spectrum Watermarking Schemes in the
Wavelet Domain Using HVS Characteristics.
International Journal of Information Science and
Technology, 5(2):1~4.
Huang Hui-Yu, Yang Cheng-Han, H. Wen-Hsing, 2010. A
Video watermarking Technique Based on Pseudo-3-D
DCT and Quantization Index Modulation. Transaction
on Information in Forensics and Security, 5(4):625-
637.
Y. Liu, J. Y. Zhao, 2010. A New Video Watermarking
Algorithm Based on 1D DFT and Radon Transform.
Signal Processing, 90(2): 626~639.
H. Ling, L. Wang, F. Zou, 2011. Robust Video
Watermarking Based on Affine Invariant Regions in
the Compressed Domain. Signal Process, 91(8):
1863~1875.
S. Gujjunoori, B. B. Amberker, 2013. DCT Based
Reversible Data Embedding for MPEG-4 Video Using
HVS Characteristics. Journal of Information Security
and Applications, 18(4): 157~166.
S. Margarita, O. Gabrieda, N. Shawn, 2009. A PSO/ACO
Approach to Knowledge discovery in a pharmacyo
vigilance context. Proceedings of the 11th Annual
Conference Companion on Genetic and Evolutionary
Computation Conference. Montreal, Canada,
2679~2684.
Y. G. Wang, L. Fan, Y. Q. Lei, 2010. PSO-Based Robust
Watermarking of AVS-Encoded Video. Proceedings
of the 2010 IEEE International Conference on
Multimedia and Expo (ICME 2010). Singapore,
1647~1650.
Yuchi Liang, Gu Jihua, Liu Wei, Tao Zhi, 2006. An Image
Digital Watermark Technique Based on Digital
Holography and Discrete Cosine Transform. Journal
of Optics, 26(3): 355~361.
N. Takai, Y. Mifune, 2002. Digital Watermarking by a
Holographic Technique. Applied Optics, 41(5):
865~873.
Zhang Wen-quan, Zhou Nan-run, 2012. Double-color
Image Encryption Based on Discrete Fractional
Random Transform. Journal of
Electronics&Information Technology, 34(7):
1727~1734.
Video Watermarking Algorithm based on Combined Transformation and Digital Holography
385
