University of Ioannina
Μπεϋζιανές μέθοδοι ανάλυσης και επεξεργασίας βιοϊατρικού σήματος και εικόνας
Abstract
dc:descriptionThis thesis is focused on the study and the development of intelligent methods for the processing of biomedical signal and image. Biomedical signals belong to the class of sequential data, i.e. data that are evolved in time or space. Their structure is complex and can be obtained in serial or batch mode. Finally, biomedical signals contain hidden characteristics and their detection constitutes a difficult task. All the above properties although bring some serious obstacles during the study of these signals, are issues of great research interest and challenges, in the sense of becoming the seeds of building effective and innovative mechanisms and tools for biomedical data analysis. Moreover, the necessity of these methods is further amplified with the fact that biomedical signals belongs to the kind of data from the real world applications. Under these prism, analyzing these pieces of information may significantly affect the human life, improve the understanding of the human body, as well as may become a light to the discovery of new perceptions and achievements within the medical world. The scope of this thesis is to study and present powerful statistical models that incorporate various interesting properties of biomedical signals, such as spatial and temporal correlations between, their time-varying nature, and their environment, in order to achieve the improvement of the fidelity of analysis and the decision making procedure. A desired feature of the models that are presented thought this thesis is to describe the signal with a single and less complex, but powerful and efficient, formulation in a way of increasing their generalization capabilities. One such representation is the sparse representation, which constitutes a modern tendency to the statistical data analysis community with many applications to several others fields, such the Biomedical Engineering, Biology, Machine Learning etc. Variations of the generalized linear regression model and the state - space models, such as the Kalman Filter, are the main stochastic models that are presented for analyzing electroencephalograms, and time series from the heart and from functional Magnetic Resonance Imaging.
Degree
thesis:*- Grantor dc:publisher
- University of Ioannina
- Year dc:date
- 2010
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Oikonomou, Vangelis
- Οικονόμου, Ευάγγελος
Subjects
dc:subject × 16- Μπεϋζιανές μέθοδοι
- Επεξεργασία βιοιατρικού σήματος και εικόνας
- Εικόνες λειτουργικής μαγνητικής τομογραφίας
- Ηλεκτροεγκεφαλογράφημα
- Προκλητά δυναμικά
- Μεταβολή καρδιακού ρυθμού
- Bayesian methods
- Biomedical signal and image processing
- Functional magnetic resonance imaging (FMRI)
- Electroencephalogram (EEG)
- Event related potentials (ERP)
- Heart rate variability
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/22264