This project aims at studying innovative techniques for protecting commercial products against counterfeiting. Two kinds of products will be considered within the project: textile fibers and fabrics on one hand and manufactured goods made of thermoplastic or including a thermoplastic substrate on the other hand. The protection will be carried out by embedding identification information directly on material surfaces, using micro and nanotechnologies (coating). The embedded identifiers are keys to databases allowing the traceability of products from factories to final customers.
In the textile case, the information is encoded during the coating process, which enables to vary the thickness of the liquid adhering to the fiber. If one can impose the diameter all along the fibers, it is possible to generate a modulated 3D surface, which actually contains a 1D information (the modulated diameter). The function representing the diameter as a function of the length of the fiber is the signal that carries the identification information.
This signal processing part of the project will study the shaping of this signal with regards to the operating conditions and the secrecy of the same signal. The control of the thickness will have a limited precision radially and on the length of the fibers. It is possible to study this problem as a communication problem, where the bandwidth of the channel is limited by the precision of the coating process. The secrecy of the signal will be carried out by the use of techniques inspired by classical theory from watermarking, which must be extended to this particular problem. As a matter of fact, the information that must be encoded through coating must be secret to guarantee authentication. The best way to do it is to use a cryptographic key to generate the signal. No one else than the owner of the key can create it. The protection against copy is brought by the complexity and the cost of the equipment necessary to realize the coating. The major challenges regarding watermarking are the use of the redundancy (spread spectrum, error correcting codes, repetition on the same fiber, repetition on several fibers) as a function of the bandwidth the required robustness (irregularity due to coating or to modification of the thickness during the lifetime of the textile).
In the case of plastic-based products, we also intend to use coating techniques to hide tracing information, but on a 2D surface instead. The information is encoded in the position of liquid thickness peaks in a given pattern. The protection will be ensured by replication of this pattern at several locations on the product. The retrieval of these patterns will be studied. This time, the same kind of signal processing study will be performed on a 2D sampled signal or possibly on a 3D volumetric surface. The study will therefore include a texture and morphology analysis, and in the 3D case, the study of watermarking for MeshGrid represented surfaces.
The physical watermarking techniques developed in this project will be decoded by specific reading devices. In this project, we intend to study their design subject to industrial considerations.
Last but not least this project will address the standardization issue at different level, with possible actions at the European level in active groups such as CEN, EAN or ETSI. The first issue is the standard definition of the identifiers, which is needed for interoperability reasons. The delivery of the identifiers must make use of trusted delivery protocols as well as certification centers that will bring trust in the IDs deliverance.
Finally, since the project is targeting mass markets, there is a requirement to study all means to decrease the final costs of the system. There should be pieces of this system (embedding and reading devices) that will have to be reproduced at a wide scale. The specifications of these devices should therefore be fixed and identical for all manufacturers for interoperability purposes. This could be made possible through the emergence of a standard.

Runtime: 2002 - 2005