You are on page 1of 4

Multimedia Copyright Protection Platform Demonstrator

Miguel Soriano, Marcel Fernandez, Elisa Sayrol, Joan Tomas, Joan Casanellas,
Josep Pegueroles, and Juan Hernndez-Serrano
Escola Tecnica Superior dEnginyeria de Telecomunicacio de Barcelona (ETSETB),
C/ Jordi Girona 1 i 3, Campus Nord, UPC, 08034 Barcelona, Spain
{soriano, marcelf, josep, jserrano}@entel.upc.es
elisa@tsc.upc.es

Abstract. The work presented in this paper consists in the development of a


portable platform to protect the copyright and distribution rights of digital
contents, and empirically demonstrate the capacity of several marking and
tracing algorithms. This platform is used to verify, at a practical level, the
strength properties of digital watermarking and fingerprinting marks. Initially,
two watermarking algorithms, one based on spread-spectrum techniques and the
other based on QIM (Quantization Index Modulation), have been implemented.
Moreover, we use these watermarking algorithms to embed a fingerprinting
code, based on code concatenation, equipped with an efficient tracing
algorithm. In this paper we focus on the implementation issues of the Javabased platform, that consists of three main packages that are fully described.

1 Introduction
The distribution and playback of digital images and other multimedia products is easy
and fast. Thus, its processing in order to achieve satisfactory copyright protection is a
challenging problem for the research community. Encrypting the data only offers
protection as long as the data remains encrypted, since once an authorized but
fraudulent user decrypts it, nothing stops him from redistributing the data without
having to worry about being caught. A watermarking scheme [1,2], which embeds
some owner information (mark) into host images, is regarded as a possible solution to
this problem. Nevertheless, digital watermarking is not strong enough to offer
protection against illegal distributors. In this environment, digital fingerprinting
techniques [3] provide a good solution to dissuade illegal copying. To make such
distribution systems work securely, the embedded marks in those systems must be
resistant to powerful attacks such as common image processing operations, lossy
image compression, geometric transforms, combination addition of random noise
(errors) and/or collusion attacks.
The work presented in this paper consists in the development of an empirical and
portable platform with the following objectives:

To offer a practical platform where digital video can be protected against


dishonest users.
To verify at a practical level the strength properties of digital watermarking
and fingerprinting marks.

P. Herrmann et al. (Eds.): iTrust 2005, LNCS 3477, pp. 411 414, 2005.
Springer-Verlag Berlin Heidelberg 2005

412

M. Soriano et al.

To compare different watermarking algorithms. In this first version, spreadspectrum and QIM (Quantization Index Modulation) based watermarking
algorithms have been implemented.
To offer a platform over which we can embed different digital fingerprinting
codes, and verify its behaviour. In the first version, we have developed the
proposal presented in [3]

An entire platform for digital video has been developed. The MPEG2 video format
has been selected since the necessary Java libraries are available and already
developed. The following section shortly describes the implementation issues. The
platform front-end is showed in section 3. Finally, section 4 is devoted to some
conclusions.

2 Platform Implementation
The platform consists of five different blocks, regarding its functionality: Web Server,
MuxDemux Server, MarkAudio Server, MarkVideo Server and Mark Generator.
Figure 1 illustrates the platform structure and the process flow to mark the multimedia
contents that should be protected.
3
1

VideoMarker Server
8

Web Server
(Tomcat)

MuxDemux
Server

4
5

Mark
Generator

AudioMarkerServer

Fig. 1. Platform structure

As example of the marking operation system, the MPEG2 copyright protection


process is detailed. The overall procedure consists on the following steps: a)
demultiplexing the MPEG-2 sequence into video and audio bitstreams, b) extraction
of DCT coefficients from the sequence, and c) the embedding of a mark (only for
traceability purposes) according to the chosen method (spread spectrum technique or
QIM). The coefficients are then put back into the sequence and multiplexed with the
audio sequence.
To achieve the functional objectives in an efficient manner, the following features
have been established:

Web interface: a common implementation with server technology based on


servlets 2.4 on top of Tomcat has been adopted. Each time that a client
generates a request the corresponding servlet establishes a socket connection
to the MuxDemux server.
Modularity: The system is designed to be highly modular so to allow in a
simple manner: the interaction, the addition of new functionalities and the
reuse of the implemented classes. When a bitstream is analyzed, all the

Multimedia Copyright Protection Platform Demonstrator

413

MPEG2 hierarchy is identified and properly encapsulated into the


corresponding objects.
Execution time minimization: the MPEG2 hierarchy objects are fully
stored in memory. Once loaded in memory, the marking process and the
sequence reconstruction is efficiently done.
Platform portability: all the system has been developed using Java.
Scalability: the modular design allows incorporating new watermarking and
fingerprinting algorithms. Moreover, the system can be easily modified and
adapted to any number of concurrent requests.
Design robustness: the system has been designed in a way that prevents
from modifying essential values that are necessary to recompose the MPEG2
sequence.

Figure 2 shows the sequential steps during the marking process.

Fig. 2. General marking process

Some difficulties have been tackled during the implementation process:


a)

No native Java libraries exist that allow an easy access to the DCT
coefficients.
b) The mark embedding should not alter the sequence quality considering the
Human Visual System (HVS,) this is why the quantization of the DCT
coefficients has to produce the minimum possible loss. So, the Java rounding
functions have been slightly modified to avoid additional losses and to
ensure a video sequence with the required robustness and imperceptibility
properties.
c) To improve the codeword search during the run-length decoding, the
codewords are ordered as a binary tree.
d) The minimum working size in Java is a byte. So, some tools have been
developed to read and write at the bit level (the minimum working size in
MPEG2).
e) The MPEG2 standard supports a large quantity of user extensions. To have a
complete implementation, the Abstract Factory model has been utilized.

3 User Platform Front-End


The platform can be accessed and tested through the following web site:
http://isg.upc.es/ongoing.html

414

M. Soriano et al.

When a user accesses the platform, he/she has to select the kind of document
(image, MPEG2 file, video bitstream ) and the desired service (watermarking,
fingerprinting ). Later the user has to choose the corresponding algorithms and their
parameters if necessary. Next, the user uploads the document or selects one from the
platform gallery. Finally the user can download the marked document.

Fig. 3. User front-end view. A) Sequence format and algorithms selection. B)Chose sequence
and configure algorithm parameters

4 Conclusions
This paper presents the implementation of a portable, modular and robust platform,
that can be used to test watermarking and fingerprinting algorithms for video
sequences. The associated drawbacks encountered during the process of
implementing the platform are discussed. Classical spread spectrum techniques have
been used for watermarking the DCT coefficients. Also, an algorithm based on dither
modulation-quantization index modulation techniques is introduced. At the present
moment spread-spectrum-based techniques have shown a better performance,
however DM-QIM techniques offer many possibilities still to be fully exploited. With
respect to fingerprinting, we use of a 2-secure fingerprinting code based on code
concatenation.

References
1. B. Chen and G.W. Wornell. Quantization index modulation: A class of provably good
methods for digital watermarking and information embedding. IEEE Trans. Inform.
Theory, 47:14231443, 2001.
2. I. Cox, J. Killian, T. Leighton, and T. Shamoon. Secure spread spectrum watermarking for
multimedia. IEEE Trans. Img. Process., 6(12):16731687, 1997.
3. M. Fernandez and M. Soriano. Fingerprinting concatenated codes with efficient
identification. In Information Security ConferenceISC, LNCS, volume 2433, pages 459
470, 2002.

You might also like