content protection has recently become an important issue because of
insufficient cognizance of intellectual property. Watermarking is one possible
method to protect digital assets, and the technology of watermarking has
extended its applications from copyright protection to content indexing, secret
communication, fingerprinting and many others.
The proposed method is a hiding biometric watermark that uses
digital video as a cover file. The recipient needs only process with required
steps in order to retrieve the watermark data. The idea of proposed method is
based on hiding the watermark in audio partition of video file instead of video’s
image. Also use multiple frequency domains to hide the biometric watermark data
using chaotic stream as key for encrypting the watermark and choose location for hiding. The
performance of the proposed algorithm is estimated by used subjective and
objective testes (SNR, PSNR and MSE) with applying simple attack that may attack
the cover file.
Experimental result of the algorithm shows good recovering of watermark
code which is virtually undetectable within video file.
Keywords: video watermarking, DCT, DWT, Biometric system, chaotic.
Currently, Internet and digital media are getting more and more
popular. So, the requirements of secure transmission of data also increased.
Various good techniques are proposed and already taken into practice 1. One
of these techniques is the watermark. A watermark is a digital code permanently
embedded into the digital cover content i.e. text, audio or video sequence 2.
Watermarking method can be described in the following process.
First, copyright data are abstracted as watermarks and cast onto multimedia
carriers by various embedding algorithms. Then, the carriers are distributed
via the computer network or digital storage. If necessary, the carriers are
processed to detect the existence of a watermark or to extract watermark bits
for different purposes3.
Generally, a practical watermarking system embeds some copyright
information into the host data as a proof of rightful ownership and must meet
requirements: Obviously, Robustness, Imperceptibility, Capacity, and Security 4.
Different digital video watermarking algorithms have been
proposed. Video watermarking techniques are classified according to their
working domain. Some techniques embed watermark in the spatial domain by
modifying the pixel values in each frame extracted from the video. These methods
are not robust to attacks and common signal distortions. In contrast, other
techniques embed the watermark in the frequency domain, which are comparatively
more robust to distortions2.
Digital video is a sequence or collection of consecutive still
images merging with audio. A watermark can carry any information you can
imagine but the amount of the information is limited. The more information a
watermark carries the more vulnerable that information is. Anyway, the amount is absolutely limited by the size
of particular video sequence2.
II. What is biometrics?
Biometrics, refers to authentication based on his or her
physiological or behavioral characteristics and its capability to distinguish
authorized and an unauthorized person. Since biometric characteristics are
distinctive as they cannot be forgotten or lost, the identification person has
to be present physically 56.
Among all biometrics such as fingerprint, facial thermogram, hand
geometry, face, hand thermogram, iris, retina, voice, signature etc.,
Iris-based identification is one of the most mature and proven technique. Iris
is colored part of eye as shown in Fig. 1. A
person’s two eye iris have different iris pattern, two identical twins also
have different iris patterns because iris has many feature which distinguish
one iris from another. Primary visible characteristic is the trabecular
meshwork. Iris is not subject to the effects of aging which means it remains in
a stable form from about age of one until death. Furthermore, iris recognition
systems can be non-invasive to their user. The use of glasses or contact lenses
has little effect on the representation of the iris and hence does not
interfere with the recognition technology57.
The chaotic signals are like noise signals but they are completely
certain, that is if we have the primary quantities and the drawn function, the
exact amount will be reproduced. The advantages of this signal are as follows 8:
The sensitivity to the
This means a minor change in primary amount will cause significant
difference in subsequent measures. It means if we have a little change in the
signal amount, the final signal will be completely different.
The apparently accidental
In comparison with productive accidental natural number
in which the range of the numbers cannot be produced again, the technique used for producing the accidental number
in algorithm based on the chaotic function will prepare the ground that if we
have the primary quantities and the drown function, we can produce the number
III. The deterministic work
As the chaotic functions were the accidental manifest, they are
completely similar. It means as we have the drawn function and the primary
quantities, we can produce and re produce sets of numbers which seemingly have
no system and order. One of the most famous signals which has chaotic features
is shown in (1), and it is known as the Logistic Map signal,
=rXn (B-xn) (1)
in which the Xn will get the numbers between 0,1. The signal
shows three different chaotic features in three different ranges on the
division of r parameter of which the
signal feature will be the best by considering X0 =0.3.
if r 0,3, then the signal feature in the first 10 repetition
show some chaos and after that it was fixed , Fig. 2 (a)910
– if r 3, 3.57, then the
signal feature in the first 20 repetition show some chaos and after that it was
fixed, Fig. 2(b),
– if r 3.57,4, then the
signal feature is completely chaotic , Fig. 2(c)
According to the given description and research requirements for
complete chaotic feature for video watermarking, the logistic map chaotic
signal with primary value X0=0.3 and r ? 3.57, 4 are used9.
The related Works
A lot of video watermarking algorithms have been proposed in the
literature employed either in spatial or frequency domain. One of these methods
was proposed by Mobasseri (2000), who suggest a spatial domain watermarking
scheme for compressed videos. Where Hong et al (2001) have proposed DWT based
algorithm in which middle frequencies are modified and a flag is generated for
extraction process. In the extraction process, another flag is generated from
watermarked image and compared with the original flag. In this algorithm,
instead of taking watermark image, authors have used generated flag as
watermark. Liu et al (2002) have proposed a wavelet transform-based video
watermarking scheme where multiple information bits are embedded into
uncompressed video sequences. Embedding is done in LL sub-band, reducing error
probabilities of detection via BHC code. Ge et al (2003) have presented a novel
adaptive approach to video watermarking. They take full advantage Wavelet
packet transform-based robust video watermarking technique 373 of both
intra-frame and inter-frame information of video content to guarantee the
perceptual invisibility and robustness of the watermark. Tsai & Chang
(2004) have proposed a novel watermarking scheme for a compressed video
sequence via VLC decoding and VLC code substitution. Zhong & Huang (2006)
have presented video watermarking based on spread-spectrum techniques to
improve watermarking robustness. Mirza et al (2007) have proposed a video
watermarking scheme based on Principal Component Analysis 4.
The proposed method
As we know video file format contain major two part of multimedia
types: image and audio. It is generated by mixing the two kinds of multimedia
types. The proposed method differs from the typical watermarking scheme. It is
based on hiding watermark data in video’s audio part instead of image one.
Digital watermark is divided into two categories: spatial domain
watermarking technique and frequency domain watermarking techniques. The
spatial domain methods embed watermark by modifying directly some values of
video file. The frequency domain methods will be better to determine perception
criterion so as to embed the watermark well 3. Therefore the proposed
algorithm used frequency domain to hide watermark data and in order to achieve
more security multiple type of frequency domains with chaotic key are used.
In the proposed method, the
watermark is based on biometrics (exactly on iris) to generate the watermarking
code. The following sections discuss the proposed video Watermarking in details.
The proposed algorithm of embedding
The proposed algorithm can be divided into two basic parts: generating
the biometric watermark code and hiding it in video file data using chaotic key.
Generating the biometric watermarking code:
Iris region consists of two circles: one for iris sclera boundary
and another for iris pupil boundary. To isolate actual iris region in eye
image, segmentation is required. To have segmentation, edge detection, circle
detection, eyelid detection are required. Various methods for edge detection
are available. Here, canny edge detection is used to find edges and Hough
transform to find iris and pupil boundaries from the image. CASIA iris image database is
used for experimentation. Iris image must be available in sender and receiver
sides. For more security the watermark is encrypted using chaotic key.
The proposed algorithm of generating the bio-watermarking code is
explained in the following steps:
Input: Iris image.
Output: Encrypted bio-watermarking code.
data which is laying under pupil circle.
edge detection using canny filter.
iris data using the generated chaotic key.
Fig. 3 shows the flowcharts of generating the bio-watermark code.
Embedding the watermark in video file using chaotic key:
Input: Video file,
Output: Watermarked video
video file to be cover file.
image and audio in it and consider audio part as a cover.
DWT on audio part.
DCT on resulted DWT coefficients.
the length of watermark (Len) in first 4 bytes of cover data.
chaotic key to be the index of chosen cover data .
watermark code in cover by exchanging the fourth decimal number after comma in
cover by another digit of watermark code.
this step until last digit in watermark code.
DCT inverse, then DWT inverse.
the video cover.
Fig. 4 shows the proposed algorithm of hiding the
biometric watermarking code in video file using chaotic key.
The proposed algorithm of extracting watermark code:
Input: The covered video file.
Output: Achieve video file protection or not.
the covered video file.
audio part from the covered video file.
DWT on audio part.
DCT on resulted DWT coefficients
the length (Len) of watermark from first 4 byte in cover.
chaotic key(for extracting and decryption operation).
the chaotic key to extract watermark code.
this step until reaching the length of watermark code.
Decrypt the extracted watermark using same
Independently… Generate the iris watermark
code (origin one) by executing the steps of generating the biometric watermark
(1 to 5).
the extracted watermark with the original one. If they are identical ,video
file protection is achieved otherwise the file is not protected.
Fig.5 shows the proposed algorithm of
extracting watermark code.
application and results
A number of video
sequences have been tested using the proposed method. The bio-watermark is
extracted from the watermarked video and its robustness is checked by calculating
some famous measures.
the proposed method is applied on many iris images obtained from CASIA
database. At last the iris code is obtained and hidden in video file. Figs
6,7,8 show the experimental steps that are done on iris image to get bio-watermark
A number of measures are
applied on it to make sure that the proposed algorithm is strong enough to
carry the watermark safely. Table I. explain the results of applying standard measures
(Correlation, SNR,PSNR and MSE) to the proposed
table I. the
results of applying standard measures to proposed algorithm
The watermarked video was attacked by simple
types of watermarking attacks. These attacks attempt to damage the embedded
watermark by modifications of the whole cover without any effort to identify
and isolate the watermark 1112. Adding white noise (Gaussian noise) is applied
to the video cover resulting from the proposed algorithm. Fig. 9 shows the
effect of adding Gaussian noise to the video cover file with different signal
to noise ratio values. While Table II. explains the output results of adding Gaussian
noise to the video cover .
Table II. The
output result of adding gaussian noise to the embedded watermark
The paper propose an
efficient method to embed abiometric watermarking in video file. It make use of
two powerful mathematical transforms:
DWT and DCT and applied them on the audio part of video file format
instead of video’s images. The proposed method use the chaotic sequence in
order to find a video file locations in order to hide bio-watermark on the one
hand and the sequence is used to encrypt
and decrypt the bio-watermark data on the other.
applying the proposed algorithm, the similarity between the original watermark
and the extracted watermark from video files is measured using correlation,
SNR, PSNR and MSE. Also measures are applied on attacked video file using
correlation and MSE. The experimental results show their robustness against
noise adding; very low noise watermark with expectable SNR values. The obtained
results give to the proposed algorithm high
performance with robustness in watermarking application in order to achieve
protection to any video file.
1. Bhaumik Arup , Choi Minkyu ,Robles Rosslin J. and Balitanes
Maricel O. ,” Data Hiding in Video”, International Journal of
Database Theory and Application, Vol. 2, No. 2, June 2009, p 9-16 .
2. Hood Ankita A. and Janwe
N. J. ,”Robust Video Watermarking Techniques and Attacks on
Watermark – A Review”, International Journal of Computer Trends and
Technology- volume4 Issue1 ,2013, p30-34.
3. Faragallah Osama S., “Efficient video watermarking
based on singular value decomposition in the discrete wavelet transform
domain”, International Journal of Electronics and Communications (AE?) , Int.
J. Electron. Commun. (AE?) 67 , 2013 , p189– 196 .
4. Bhatnagar Gaurav and Raman
Balasubrmanian, “Wavelet packet transform-based robust video watermarking
technique”, Indian Academy of Sciences , Sadhana Vol. 37, Part 3, 2012, p 371–388.
5. Al-Gurairi Maha Abdul-Rhman Hasso,” Biometric
Identification Based on Improved Iris Recognition Techniques”, A Ph. D.
Thesis Submitted to The Council of the College of Computer and Mathematical
Sciences, University of Mosul ,2006.
6. Waghmare L.M. and Roselin
Vanaja, ” Iris Texture Analysis for Security Systems” , International
Journal of Computer Applications (0975 – 8887) Volume 64– No.22, 2013, p37-44.
7. Dhavale Sunita V. ,”
DWT and DCT based Robust Iris Feature Extraction and Recognition Algorithm for
Biometric Personal Identification “, International Journal of Computer
Applications (0975– 8887) Volume 40– No.7, 2012, p 33-37.
8 Enayatifar R. , Mahmoudi
F. and Mirzaei K.,” Using the chaotic map in image steganography.
International Conference on Information Management and Engineering, 2009 ,p 491-495.
9. Saeed Melad J., ” A New technique based on
chaotic steganography and encryption text in DCT domain for color image”, Journal of
Engineering Science and Technology Vol. 8, No. 5 , 2013, p508 – 520 .
10.Ahmed H.E., Kalash, H.M. and Farag Allah, O.S., “An
efficient chaos-based feedback stream cipher (ECBFSC) for image encryption and
decryption”, Informatica, 31(1), 2007 ,p 121-129.
11. Ali Dujan Basheer Taha ,
“Digital Image Watermarking Techniques For Copyright Protection”, A
Ph. D. Thesis Submitted to The Council
of the College of Computer Sciences & Mathematics , University of Mosul. ,
12.Zlomek Martin, ” Video Watermarking”,
master thesis submitted to Department of Software and Computer Science Education, Charles University in Prague , Faculty of
Mathematics and Physics, 2007.