Classification of 3-D MRI Brain Data Using Modified Maximum Uncertainty Linear Discriminant Analysis

Warning

This publication doesn't include Faculty of Economics and Administration. It includes Faculty of Medicine. Official publication website can be found on muni.cz.
Authors

JANOUŠOVÁ Eva SCHWARZ Daniel KAŠPÁREK Tomáš

Year of publication 2010
Type Article in Proceedings
Conference Proceedings of Medical Image Understanding and Analysis 2010
MU Faculty or unit

Faculty of Medicine

Citation
Web http://www2.warwick.ac.uk/fac/sci/dcs/events/miua2010/proceedings/
Field Neurology, neurosurgery, neurosciences
Keywords Classification, Principal Component Analysis, Linear Discriminant Analysis, MRI, Computational Neuroanatomy, Schizophrenia
Description Recent studies have demonstrated that diagnostics of schizophrenia based on image data is a difficult task because of extensive overlaps of brain regions distinguishing patients with schizophrenia from healthy controls and also because of the small sample size problem. An algorithm for the automatic classification of first-episode schizophrenia patients and healthy controls based on deformations and gray matter (GM) density images extracted from their MRI intensity data is introduced here. The deformations and GM density images are reduced by principal component analysis, which is here based on the covariance matrix of persons (pPCA). The reduced image data is then classified with the use of modified maximum uncertainty linear discriminant analysis (MLDA), which gives better sensitivity than original MLDA. The classification efficiency of the proposed algorithm is comparable with other state-of-art studies in the schizophrenia research.
Related projects:

You are running an old browser version. We recommend updating your browser to its latest version.