TIBM: Group comparison

Group comparison consists in highlighting the differences between two populations of individuals, with the objective to develop new knowledge in neuroscience. The problem of comparing an individual with a normal model (atlas) is also addressed in order to identify pathological areas of a given subject. Thus, work was conducted to demonstrate the advantages of multivariate statistical tests on 2nd order tensor for group study in diffusion MRI 2-BNHLxx. An approach combining dimension reduction and statistical modeling in the reduced space was proposed for the analysis of 4th order tensor, with application to the comparison of an individual with a population and the comparison of two populations 4-GRHK15. Finally, an original approach based on Gaussian Markov Random Field modeling has been proposed for comparing an individual to a normal statistical model with application to the detection of maxillofacial abnormalities 2-Fais12.

PhD thesis

• A. Bouchon, Modèles de régression multivariés pour la comparaison de populations en IRM de diffusion, september 2016
• T. Gkamas, Statistical modelling of high order tensors in diffusion weighted magnetic resonance imaging, september 2015
• G. Sfikas, Modèles statistiques non linéaires pour l'analyse de formes. Application à l'imagerie cérébrale, september 2012
• A. Belghith, Indexation de spectres HSQC et d’images IRMf appliquée à la détection de bio-marqueurs, march 2012
• F. Renard, Création et utilisation d'atlas en IRM de diffusion. Application à l'étude des troubles de la conscience, september 2011
• M. Brucher, Représentations compactes et apprentissage non supervisé de variétés non linéaires : application au traitement d’images, october 2008
• T. Vik, Modèles statistiques d'apparence non gaussiens. Application à la création d'un atlas probabiliste de perfusion cérébrale en imagerie médicale., september 2004

→ Back to Biomedical image processing page