Image, statistics and data mining: computational methods to analyse the human brain

• Grant number: 05/02899-4
• Support type: Research Grants - Young Investigators Grants
• Duration: October 01, 2006 - September 30, 2010
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal researcher: Carlos Eduardo Thomaz
• Grantee: Carlos Eduardo Thomaz
• Home Institution: Campus de São Bernardo do Campo. Centro Universitário da FEI (UNIFEI). Fundação Educacional Inaciana Padre Sabóia de Medeiros (FEI). São Bernardo do Campo , SP, Brazil
• Assoc. researchers: Edson Amaro Junior ; Geraldo Busatto Filho ; Koichi Sameshima ; Paulo Eduardo Santos ; Wagner Farid Gattaz

Abstract

The link between brain structure and function has been of interest since the beginning of brain research and has motivated several research studies all over the world. Advances in neuroscience and neuroimaging have led to an increasing recognition that certain neuroanatomical structures may be affected preferentially by particular diseases. The development of computational methods for morphological analysis of the human brain has enabled the automatic characterization and quantification of neuroanatomical differences between a reference group of images and the population under investigation. In the recent years, multivariate pattern recognition methods that analyse all voxels simultaneously have been proposed to classify and analyse morphological and anatomical structures of 3D magnetic resonance (MR) brain images. Most of these techniques have overcome the difficulty of dealing with the inherent high dimensionality of 3D MR brain image data by using pre-processed segmented images or a small number of specific features. However, an intuitive way of mapping the classification results back into the original image domain for further interpretation remains challenging. The aim of this project is to develop an integrated framework for classifying and analyzing patterns of disorders from medical images using a combination of image registration, multivariate statistics, and data mining. Our goal is to analyse all the data simultaneously rather than feature by feature. The first part of this project consists basically of studying and evaluating multivariate statistical methodologies to identify the most discriminating hyper-plane separating two populations. In the second part of this project we are going to develop an intelligent agent aimed at assimilation and exploration of the knowledge contained in the results of the previous statistical methods. Automated knowledge discovery from the visual observation of images is the key motivation for the methods to be investigated in this research. We believe that such investigation provides a suitable framework for characterizing the high complexity of MR images in schizophrenia, Alzheimer, and epilepsy. (AU)

Scientific publications

(5)

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

THOMAZ, CARLOS EDUARDO; GIRALDI, GILSON ANTONIO. A new ranking method for principal components analysis and its application to face image analysis. Image and Vision Computing, v. 28, n. 6, p. 902-913, JUN 2010. Web of Science Citations: 201.

SATO, JOAO RICARDO; FUJITA, ANDRE; THOMAZ, CARLOS EDUARDO; MORAIS MARTIN, MARIA DA GRACA; MOURAO-MIRANDA, JANAINA; BRAMMER, MICHAEL JOHN; AMARO JUNIOR, EDSON. Evaluating SVM and MLDA in the extraction of discriminant regions for mental state prediction. NeuroImage, v. 46, n. 1, p. 105-114, MAY 15 2009. Web of Science Citations: 33.

SATO, JOAO RICARDO; THOMAZ, CARLOS EDUARDO; CARDOSO, ELLISON FERNANDO; FUJITA, ANDRE; MORAIS MARTIN, MARIA DA GRACA; AMARO, JR., EDSON. Hyperplane navigation: A method to set individual scores in fMRI group datasets. NeuroImage, v. 42, n. 4, p. 1473-1480, OCT 1 2008. Web of Science Citations: 23.

GIRALDI‚ G.A.; RODRIGUES‚ P.S.; KITANI‚ E.C.; SATO‚ J.R.; THOMAZ‚ C.E. Statistical learning approaches for discriminant features selection. Journal of the Brazilian Computer Society, v. 14, n. 2, p. 7-22, 2008.

SANTOS‚ P.E. Reasoning about depth and motion from an observer’s viewpoint. Spatial Cognition and Computation, v. 7, n. 2, p. 133-178, 2007.