Marie-Curie Fellow
Research & Training Network
O. Picone (ESR)

Marie Curie Fellow (ESR) of the FAST project
Chalkis Institute of Technology, Greece
Contact   

• T4 Signal processing
  Task 4 focuses on MRSI signal processing proper of Gigabytes of data. The ultimate goal is reliable quantitation of metabolites. The required computing times are long and will eventually need grid-computing. Partner 6 (CIT) focuses on signal processing, neural networks, bayesian inference and pattern recognition:

  • Image processing and analysis, pattern recognition techniques for the enhancement and segmentation of MR images, including morphological filtering, coordinate logic operations, partial differential operators, wavelet descriptors
  • Investigation of MRSI scan-time reduction by means of sparse sampling. Image reconstruction using support vector machines based on wavelet analysis, bayesian neural networks as well as semi-parametric artificial neural network compared in terms of speed and reconstruction accuracy
  • MR spectrum classification and modelling, efficient unsupervised neural network algorithms investigation in clustering and modelling, probality distribution of MR spectra data based mainly on ICA as well as on PCA/ICA based Kohonen type selforganizing maps
  • New approaches, based on the compactly supported orthogonal wavelet network, investigated in terms of modelling accuracy