You are here : English versionTeamsRFAI

LIFAT - RFAI

Pattern Recognition and Image Analysis


The Pattern Recognition and Image Analysis team (in french, RFAI, for Reconnaissance des Formes et Analyse d'Images) focuses on machine learning, data mining and image processing. It conducts research on methods allowing to construct and exploit high-semantic level representations from various input data. This gathers models and algorithms at different stages in the analysis process: from low level processing (filtering, segmentation, detection of interest points, ...) to higher level processing (matching, classification, indexing, ...).

In particular, members of the team work on
  • interactive methods: integration of user input and prior knowledge into the recognition process, visual data mining embedded in virtual realty environments, ...
  • graph-based methods: construction of structured representations for matching and classification, ...
  • machine learning: incremental methods, biology-inspired algorithms, convolutional neural networks, ensemble learning, ...
  • image processing: filtering, segmentation, variational methods, discrete and continuous optimization, extraction of texture features, ...

Designed methods are used to analyze the following types of data:
  • Digitized documents
  • Real-world images and videos
  • 3D point clouds and volume medical images
  • Time series

For extra information, such as ressources and events, visit the RFAI wiki.