General information

Research for understanding spatio-temporal events detected through multiple sensors

Our goal is the interpretation and organization of information coming from distributed sensors. A sensor can be a dynamic or static imaging device, other multimedia sources, or a network of different sensors. The challenge is the evaluation, recognition and classification of events occurring at different locations or time. We place special emphasis on machine learning, data mining, human perception, geometrical optics, optimization methods and variational analysis. Our areas of interests include image and video processing, biometrical identification, connections with sensor networks and computer graphics.
 

Research objectives:

  • Classification and indexing of multimedia events
    • Stochastic models,
    • Learning methods for event analysis,
    • Self organizing distributed sensor networks and related calculations,
    • Operation optimization,
    • Event matching of multiple views,
  • Recognition and tracking of biometric features,
  • Energy optimization based image and video segmentation,
  • Automatic camera registration,
  • Optics-oriented image understanding,
  • Cognitive vision problems,
  • Compression,
  • Identification, recognition,
  • Terrestrial and aerial LIDAR data analysis,
  • Remote sensing, long term change detection and tracking in aerial images,
  • Detection and removal of artifacts from still images and videos: shadow regions, reflections, undesired objects,
  • Content-based image and video analysis, indexing, search and retrieval.
  • Multi-target detection and tracking.

Recent projects:

See Projects/Current for an up to date list of on-going research projects.

  • Multi sEnsor Data fusion grid for Urban Situational Awareness (MEDUSA, EDA JIP-FP)
  • Array Passive ISAR Adaptive Processing (EDA)
  • THIS (Transport Hub Intelligent video System) (EU FP)
  • Video indexing and classification through learning
  • Multimedia Understanding through Semantics, Computation and Learning (MUSCLE, EU NoE)
  • Digital restoration of archive movie films (DIMORF)
  • Event-driven surveillance with multicamera systems
  • Judicial Management by Digital Libraries Semantics (EU - FP7)