International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 4 Issue: 4

ISSN: 2321-8169
578 - 583

________________________________________________________________________________________________________

Watermarking for Security in Database

Prof. Manoj Dhande

Aishwarya Kotyankar

Nidhi Mannadiar

Department of Computer Engineering of

Shah and Anchor Kutchhi Engineering
College, Chembur, University of
Mumbai, India

Department of Computer Engineering

Shah and Anchor Kutchhi Engineering
College, Chembur
Mumbai, India

Department of Computer Engineering

Shah and Anchor Kutchhi Engineering
College, Chembur
Mumbai, India

manoj.dhande@gmail.com

aishwaryakotyankar.2016@gmail.com

nidhimannadiar.2016@gmail.com

Prajkta Jagtap
Department of Computer Engineering
Shah and Anchor Kutchhi Engineering College, Chembur
Mumbai, India
prajktajagtap.2016@gmail.com
Abstract: Watermarking technology is used to embed copyright information in objects such as images, audio, video and databases. The
increasing use of relational database systems in many real-life applications creates the need for database watermarking systems for protection of
database. As a result, watermarking relational database system deals with the legal issue of copyright protection of database system. There are
different types of databases like, Numerical and Categorical Databases. Working with numerical data is easier as compared to categorical
databases which is much harder to work with. This report addresses a unique, robust copyright protection scheme for Relational Database.
Watermark (Characteristic code) is a binary string calculated through the characteristic operation on the original database. A watermark is called
robust if it resists a designated class of transformations. Robust watermarks may be used in copy protection applications to carry copy and access
control information. The algorithm correlates characteristics according to the content of the databases, which can resist invertibility attack
efficiently.Invertibility attack on database is being considered in this paper.
Keywords: Digital Watermarking, Authentication, Relational Databases, Invertibility Attack, Relating Characteristic.

__________________________________________________*****_________________________________________________

I.

INTRODUCTION

Digital multimedia watermarking embeds the copyright

information in digital objects like images, audio and video.
Digital watermarking embeds information into a digital
signal in a way that is difficult to remove. Where, the signal
can be an audio, pictures or video. If the signal is copied,
then the information is also carried in the copy. A signal
may carry several different watermarks at the same time.

Figure 1. Dual Watermarking Process

Applications:

A. Types of Watermarking:
1. Visible Watermarking
In visible watermarking, the information is visible in the
picture or video. The information is text or a logo which
identifies the owner of the media. When a television
broadcaster adds its logo to the comer of transmitted video,
this is also a visible watermark.
2.

3.

B .Classification Model of Database Watermarking:

1.

Invisible Watermarking

In invisible watermarking, information is added as digital

data to audio, picture or video, but it cannot be perceived as
such (although it may be possible to detect that some
amount of information is hidden). Database watermarking
comes under this category.
Dual Watermarking:

It is a combination of both visible and invisible

watermarking.

Source Tracking.
Broadcast Monitoring.
Content protection for audio and video content.
Forensics and piracy deterrence.
Communication of ownership and copyright.
Document and image security.

2.

3.

Data Type: Different schemes are designed for

watermarking different types of data, including
numerical data and categorical data.
Distortion to underlying data: While some
watermarking
schemes
inevitably introduce
distortions/errors to the underlying data, others are
distortion-free.
Sensitivity to database attacks: A watermarking
scheme can be either robust or fragile to database
attacks. A scheme is robust (fragile, respectively) if
it is difficult to make an embedded watermark
undetectable (unchanged, respectively) in database
578

IJRITCC | April 2016, Available @ http://www.ijritcc.org

_______________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 4 Issue: 4

ISSN: 2321-8169
578 - 583

________________________________________________________________________________________________________
attacks, provided that the attacks do not degrade the
1. Generate Hash Value:
usefulness of the data significantly.
A database is embedded with a watermark using a private
4.
Watermark information: The watermark information
secret key. A new virtual primary key is generated using a
that is embedded into a database can be a single-bit
cryptographic secure one-way hash function, which consists
watermark, a multiple-bit watermark, a fingerprint,
of with a secret key (KEY) in conjunction with the database
or multiple watermarks in different watermarking
table primary key (pk). Hashing algorithm-MD-5 is used.
schemes.
5.
Verifiability: A watermark solution is said to be
2. Encode Watermark:
private if the detection of a watermark can only be
To embed mark bits. All bits will be XORed with securely
performed bysomeone who owns a secret key and
selected MSB (Most Significant Bit) of selected attribute
can only be proven once to the public (e.g., to the
among changeable candidate attributes, according to
court). After this one-time proof, the secret key is
generated key. This approach will increase mark
known to the public and the embedded watermark
imperceptibility.
can be easily destroyed by malicious users. A
watermark solution is said to be public if the
3. To apply watermark to the Database:
detection of a watermark can be publicly proven by
anyone, as many times as necessary.
Two dimensions are used to refer the image directly, and
6.
Data structure: Different watermarking schemes are
does not map the image from two dimensions to a vector. A
designed to accommodate different structural
hash function is used to generate values for a tuple. We then
information of the underlying data, including
use the generated values and use it for elements coordinates
relational databases (with or without primary keys),
and the value of pixel in this coordinates will be embedded
data cubes, streaming data, and XML data.
into the relevant tuple.
4. Minimize Variation of Data:
Altering the value of attributes is the most important point in
database during watermarking and should be noted. The
vital issue is to limit modifications in the tuples, to an
acceptable range.
5. Selection of LSB:
In this method one of the specified LSBs will be selected
and modified.The attribute will be changed as well. If
thefirst LSB is the selected bit, the value will be changed in
minimum rate, but if the chosen bit is the second or a higher
bit, the value will be changed by more than one unit.
6. Commit:
Figure 2. Watermarking Process
C.PROPOSED ALGORITHM PROCESS

All modifications will be saved into the database

by issuing Commit command, provided all of the above
steps are performed without any error.
D. Analysis of Invertibility Attack Resistance
An attacker may take following steps to find the fictitious
watermark.
(1) Choose the parameters randomly and use them to do
extraction operation on the watermarked database R_;
(2) From each tuple an index number can be computed
(index (ri) =rand (K, ri.P)). And in this way he can choose
(3) Calculate the watermarked attributes and the LSBs
(attribute index x=index (ri) mod , bit index j= index (ri)
mod );

Figure 3. Detailed Process [8]

(4) At last, extract the watermark information from the

effective bits and obtain the fictitious watermark WM_.
579

IJRITCC | April 2016, Available @ http://www.ijritcc.org

_______________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 4 Issue: 4

ISSN: 2321-8169
578 - 583

________________________________________________________________________________________________________
Then he can forge the original database according to the
same with the legitimate owners by an attacker is shown as
fictitious watermark WM_ by using the same parameters to
formula (6):
fictitious watermark.
P =1
1000 1000 4 50
= 0.5 108 (6)
Once a forged original database Ris obtained, the attacker
So it can be concluded that it is impossible for an attacker to
may declare his copyright of database R_ by providing
select all the parameters same with those of the legitimate
forged original database R, watermark WM_ and
owner.
watermarked database R_.
At the same time, we can conclude from Table 3 that
different parameters have different effect on watermark
To confirm the copyright, a legitimate owner can providethe
detection. The effect of for watermark detection is
original database R, watermark WM, watermarkeddatabase
stronger.
R_. According to the proposed algorithm, characteristic code
for R and Rcan be computed respectively:
III.PROPOSED SYSTEM

All different 0.49 restore the watermarked database. The

original bit before embedding is either 1 or 0. He can
randomly change the data at the calculated location with 0 or
1, according to the O(R) = WM, O (R) = WM
Calculate the similarity 1 and 2:
1 = sim (WM, WM), 2 = sim (WM_, WM)
Since the fictitious watermark WM_ is randomly found,
The similarity 2 is relatively low while the similarity 1 is
higher.

To identify the tuples in the relational database and

embedding a watermark in them in spatial domain. This
watermarked relational database should not be vulnerable to
attacks like invertibility attack. The end result will be a
relational database with copyright protection that cannot be
tampered with or modified without the author s permission
or access rights. The used method will insert text and/or
image watermark, detect the important tuples that need to be
watermarked so that the entire database gets protected
against unauthorized access. We also perform encryption
and decryption of the database to enhance security. We also
provide a query tuner to optimize the queries. We perform
fragmentation and decomposition where we propose a
futuristic view to watermark the fragmented and
decomposed database.
1.The Query tuner optimizes a query by prompting the
appropriate solution.

In the experiment, an invertibility attack was simulated.

The similarity 1 and 2 were calculated respectively. Inthe
experiment, 1 = 0.99, greater than the given threshold
. We also calculated the values of 2 according to different
values of parameters, shown as Table 3.
Table 3 shows that if one of the four parameters K, ,
, selected by the attacker is different with the
legitimateowner, the similarity will be much less than the
giventhreshold in watermark detection. The probability of
selecting all the parameters same with the legitimate
owners by an attacker is shown as formula (5):
P =1
K
(5)Assuming K [1, 1000], [1,1000], and [1,4], there
are 50 numerical attributes to be embedded watermark. The
probability of selecting all the parameters which are the

Figure 1: Query Prompt

2.The database is encrypted using a key and the same
key is used to decrypt the database
2.1 Data Encryption Standard or DES Algorithm:
1.
2.
3.

Firstly, we need to process the key.

Process a 64-bit data block.
Get a 64-bit data block
580

IJRITCC | April 2016, Available @ http://www.ijritcc.org

_______________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 4 Issue: 4

ISSN: 2321-8169
578 - 583

________________________________________________________________________________________________________
4. Perform the following permutation on the data
block.
5. Split the block into two halves. The first 32 bits are
called L[0], and the last 32 bits are called R[0].
6. Apply the 16 subkeys to the data block.
7. Expand the 32-bit R[i-1] into 48 bits according to
the bit-selection
8. Exclusive-or E(R[i-1]) with K[i].
9. Break E(R[i-1]) xor K[i] into eight 6-bit blocks
10. Substitute the values found in the S-boxes for all
Figure 2.3: Encrypted Database
B[j].
11. until all 8 blocks have been replaced.
12. Repeat the same for L[i-1]
13. Perform the permutation on the block R[16] L[16].
2.2 Decryption Algorithm :
1.
2.
3.

To decrypt, use the same process as encryption

with one change.
Use the keys K[i] in reverse order.
Instead of applying K[1] for the first iteration,
apply K[16], and then K[15] for the second, on
down to K[1].

Figure 2.4: Decryption of Database

Figure 2.5: Decrypted Database

3.Horizontal and vertical Fragmentation .

Figure: 2.1 AES encryption and Decryption process

Figure 3.1 Horizontal Fragmentation

Figure 2.2: Encryption of Database

581
IJRITCC | April 2016, Available @ http://www.ijritcc.org

_______________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 4 Issue: 4

ISSN: 2321-8169
578 - 583

________________________________________________________________________________________________________

Figure 3.2 Vertical Fragmentation

4. Decomposition provides .mdl file which shows
cardinality between entities of schema.
Figure 5.3 : Database Watermarking using image

Figure 4 : Schema Decomposition

5.Watermark.wes
file
contains
the
database
watermarked by text and watermarking.wes file contains
database watermarked by image.

Figure 5.1 : Database Watermarking using text

Figure 5.2 : Database Watermarked with text

Figure 5.4 : Database Watermarked with image

IV.CONCLUSION:
The algorithm presented in this report is a new, novel and
robust watermarking technique for watermarking relational
database. The implementation results clearly show that the
use of LSB bit replacement technique proves for an efficient
way to watermark relational database for providing
copyright protection to the database owner. We also encrypt
and decrypt the database using DES algorithm. This feature
provides for enhanced security. The query tuner provides for
optimized querying of the relational database. The
fragmentation and decomposition of database provides for a
futuristic scope wherein we propose the watermarking
process to not be limited to embedding the watermark to the
entire database but also should provide for embedding the
watermark in the desired fragments of the database as per
user convenience and desirability. Results and relative
analysis show that database is authenticated by embedding
watermark text or image and its security is enhanced by
encryption processes .The basic system encompasses of a
query tuner, Schema decomposer, a provision to watermark
the database using text and image .Database can be
encrypted and decrypted in the system. The main advantage
of this method is that it is robust and can withstand various
attacks. The system therefore provides for an efficient
system that confers security to database with authentication
582

IJRITCC | April 2016, Available @ http://www.ijritcc.org

_______________________________________________________________________________________

International Journal on Recent and Innovation Trends in Computing and Communication

Volume: 4 Issue: 4

ISSN: 2321-8169
578 - 583

________________________________________________________________________________________________________
.In the experiments, database is used as the host, whereas
text and .jpeg images are used as watermark image.
V. ACKNOWLEDGEMENT
A Paper is a teamwork and reflects the contribution of
many people .We would like to thank everyone who has
contributed to this effort by sharing their time and taking
interest in our work and encouraging us all the way
through.In particular, we thank our guide Mr. Manoj
Dhande, for helping in our project. We thank him for his
guidance, support and words of encouragement throughout
the time of our project.

VI.REFERENCES
[1] R. Agrawal, P. J. Hass and J. Kiernan, Watermarking
relationaldata: framework, algorithms and analysis,
Proceedingsof the 28th VLDB Conference, 157-169,
2003.
[2] Y. Zhang, D. Zhao, and D. Li, Digital watermarking for
relational databases, Computer engineering and
application, No 25, 193-195, 2003. (In Chinese)
[3] Y. Wang, G. Zhu and S. Zhang "Research on the
Watermarking Algorithm based on Numerical Attributein
the Relational Database", IEEE, ICCSEE, pp. 363-367,
2012.
[4] M. Atallah and S. Wagstaff. Watermarking with
quadratic residues. In Proc. of IS&T/SPIE Conference on
Security and Watermarking of Multimedia Contents,
January 1999.
[5] K. Huang, M. Yue, P. Chen, Y. He and X. Chen "A
Cluster-based Watermarking Technique for Relational
Database", IEEE, IWDTA, pp. 107-110, 2009.
[6] Atallah M, Wagstaff S (1999) Watermarking with
quadrati residues.In: Proceedings of IS&T/SPIE
conference on security and watermarking of multimedia
contents, January 1999
[7] Y. Li, H. Guo and S. Jajodia, Tamper Detection and
Localization for Categorical Data Using Fragile
Watermarks. Proceedings of the 4thACM Workshop on
Digital Rights Management, Washington DC, USA, pp.
73-82, 2004.
[8] http://www.globalcis.org/jcis/ppl//JCIS1-088017IP4.pdf.

583
IJRITCC | April 2016, Available @ http://www.ijritcc.org

_______________________________________________________________________________________