Problem :
The development of new multimedia communication applications have increased
problems linked to intellectual property or diffusion rights. In fact, with the wide
development of new digital techniques it is now possible to make perfect copies of the
digital data contents. Thus, the illegal distribution control has became quite impossible. To
try and solve these problems, lots of study on data hiding has started in the 90’s to help in
data content authentication.
In this context, digital watermarking and steganographic studies are often proposed and
solutions for specific applications have been developed. Sometimes, these data hiding
methods are combined with cryptographic methods as luring techniques.
In our study, the digital media that we are interested in is Audio. In this field the illegal
distribution has became a wide problem due to the development of exchange networks on
the Internet. In this context, no technical solutions have been found yet (excepted tribunal
actions) to control the audio legal diffusion and exchange on open networks as the
Internet.
We propose here to study and develop a specific embedding method for which the digital
media will be lured and marked in a compressed domain. By using the lured information and the data mark, we
want to obtain a watermarked readable media, containing for example the licensed
information for each legally obtained media. The lured information and the data mark
should be undissociated. The watermark embedded in the readable document should be
robust and have good properties in term of undetactability and capacity. An informed
watermarking approach will be preferred. Audio Compressed data will be considered.

Work Context :
This work belong to a RIAM Project from the ANR (National Research Agency in France)
named MEDIEVALS. The main objective of this project is for the contribution to secured
services for media exchange in networks.
This work will be done in the Signal Division of the LSS laboratory.

Candidate profile :
The post-doctoral candidate should have a large experience in signal processing and
associated tools. He must also be competent in Information theory, data quantization
methods, channel coding, vitterbi, … The candidate should also know watermarking
methods and steganography techniques.

Supervisors :
Claude Delpha, MCF UPS, claude.delpha@lss.supelec.fr,
Rémy Boyer, MCF UPS, remy.boyer@lss.supelec.fr