A
Antoine Souloumiac
Researcher at French Alternative Energies and Atomic Energy Commission
Publications - 53
Citations - 4979
Antoine Souloumiac is an academic researcher from French Alternative Energies and Atomic Energy Commission. The author has contributed to research in topics: Blind signal separation & Brain–computer interface. The author has an hindex of 16, co-authored 49 publications receiving 4752 citations. Previous affiliations of Antoine Souloumiac include Télécom ParisTech.
Papers
More filters
Journal ArticleDOI
Blind beamforming for non-gaussian signals
TL;DR: In this paper, a computationally efficient technique for blind estimation of directional vectors, based on joint diagonalization of fourth-order cumulant matrices, is presented for beamforming.
Journal ArticleDOI
Jacobi Angles for Simultaneous Diagonalization
TL;DR: This note gives the required Jacobi angles in close form for simultaneous diagonalization of several matrices.
Journal ArticleDOI
xDAWN Algorithm to Enhance Evoked Potentials: Application to Brain–Computer Interface
TL;DR: An unsupervised algorithm is proposed to enhance P300 evoked potentials by estimating spatial filters; the raw EEG signals are projected into the estimated signal subspace, and the results show that the proposed method is efficient and accurate.
Journal ArticleDOI
Nonorthogonal Joint Diagonalization by Combining Givens and Hyperbolic Rotations
TL;DR: A new algorithm for computing the nonorthogonal joint diagonalization of a set of matrices is proposed for independent component analysis and blind source separation applications and compares favorably with existing methods in terms of speed of convergence and complexity.
Journal ArticleDOI
Comparison of Eight Methods for the Estimation of the Image-Derived Input Function in Dynamic [18F]-FDG PET Human Brain Studies:
Paolo Zanotti-Fregonara,El Mostafa Fadaili,Renaud Maroy,Claude Comtat,Antoine Souloumiac,Sébastien Jan,Maria Ribeiro,Véronique Gaura,Véronique Gaura,Avner Bar-Hen,Regine Trebossen +10 more
TL;DR: Eight methods for the estimation of the image-derived input function (IDIF) in [18F]-FDG positron emission tomography (PET) dynamic brain studies were compared, and only one of the methods allowed a reliable calculation of the individual rate constants.