COLIBRI

Our laboratory supports a FIRST project : COLIBRI

Interactive kiosks are currently found in all major market segments: education, leisure, etc. They offer users a number of services that are generally incomplete, static and not truly interactive.
Since the popularization of equipment using multi-tactile technologies (iPhone, etc.), many kiosk manufacturers have been investing in the development of products to use these technologies. These companies also want to increase interactivity thanks to current (Bluetooth, RFID...) and emerging technologies (Flight Time Camera,...)
Most products on the market today are restricted to applications selected by the manufacturer and/or are sometimes very expensive. Developers of interactive kiosk applications are therefore "blocked" by proprietary development tools - often limited - linked to hardware.
Interactive kiosks are currently found in all major market segments: education, leisure, etc.
The objective of the project is to develop multimodal interactive terminals. They will, for example, be able to recognize gestures, sounds, objects, people... and interact with their environment.
The work will consist in developing a set of reusable software tools allowing rapid prototyping and easy implementation of these different tools. The information provided by the terminals will therefore be more dynamic and interactive since it will depend on the factors surrounding the terminal: objects (train ticket whose bar code is scanned), presence of people (via an RFID badge or camera), mobile devices (mobile phone via Bluetooth)...
The originality of the research consists in the development of simple tools that are independent of current multi-tactile screen manufacturers in order to enable companies that will use them to create ergonomic and ecological multi-touch interactive multi-mode equipment.

SAMIFIS

Our laboratory supports a FIRST project : SAMIFIS

SAMIFIS aims to develop an anti-spam application.
Spam is an endemic phenomenon on the Internet. It causes undesirable effects such as: network overload, mail server congestion, virus propagation, phishing. From a simple user point of view, spam is above all a daily nuisance.
Solutions already exist that allow spam filtering, separating spam from legitimate email. The methods used are numerous: from word by word analysis via regular expressions to dynamic learning via artificial intelligence. The combination of several techniques to achieve more satisfactory results than used individually. One of the added values of the project would therefore be to offer modular and adaptable tools. Aspects of the research will include:

  • artificial intelligence: learning, neural networks, vector machine support
  • algorithmics: multi-criteria research, data analysis
  • Collaboration: shared learning between machines

IMAGINE

Our laboratory supports a FIRST project : IMAGINE

IMAGINE (study of the implementation of AlGorithms for image processing in an open hardware system) aims to accelerate image processing algorithms running on computers. To achieve this, the NVIDIA GPU computing architecture (CUDA) is used, providing a very high level of computing power. The project includes the creation of graphic software modules to create applications by assembling different image processing modules and thus generate high-performance code.