Ensemble Methods in Medical Decision Making
| dc.contributor.advisor | Hajdu, András | |
| dc.contributor.author | Antal, Bálint | |
| dc.contributor.department | Informatikai tudományok doktori iskola | hu |
| dc.contributor.submitterdep | DE--TEK--Informatikai Kar -- Komputergrafika és Képfeldolgozás tanszék | |
| dc.contributor.submitterdep | DE--TEK--Informatikai Kar -- Komputergrafika és Képfeldolgozás tanszék | |
| dc.contributor.submitterdep | DE--TEK--Informatikai Kar -- Komputergrafika és Képfeldolgozás tanszék | |
| dc.date.accessioned | 2013-03-08T08:38:18Z | |
| dc.date.available | 2013-03-08T08:38:18Z | |
| dc.date.created | 2012 | hu_HU |
| dc.date.defended | 2013-03-22 | |
| dc.date.issued | 2013-03-08T08:38:18Z | |
| dc.description.abstract | Computer-aided decision support in medical problems is a prominent research area nowadays. In this PhD thesis, two approaches are shown to support the medical decisions for diabetic retinopathy (DR). This disease is one of the most common causes of blindness in the developed countries. Thus, timely and precise detection is essential for a large population. Furthermore, high reliability of the diagnosis is also desired. The first major contribution of this thesis is an approach to the early detection of DR on color fundus images. This approach aims to detect the earliest signs of DR, namely microaneurysms (MAs). Since MA detection in color fundus images is a very challenging task, we propose a novel ensemble-based framework to ensure reliable fusion of MA detection output, namely <preprocessing method, candidate extractor> ensembles. This approach proved its capabilities in exhaustive evaluation, including a competition dedicated to the comparison of MA detectors (Retinopathy Online Challenge), where it is currently ranked as first. The second major contribution of this thesis is an approach to retinal image grading based on the detailed analysis of color fundus images. Both the detected anatomical parts and presence of lesions are considered as features. We use an ensemble of machine learning classifiers for grading the retinal images based on the extracted features. As the results on a publicly available database show, a highly accurate grading system is achieved in this way. | hu_HU |
| dc.format.extent | 124 | hu_HU |
| dc.identifier.uri | http://hdl.handle.net/2437/161660 | |
| dc.language.iso | hu | hu_HU |
| dc.language.iso | en | hu_HU |
| dc.subject | orvosi képfeldolgozás | hu_HU |
| dc.subject | gépi tanulás | hu_HU |
| dc.subject.discipline | Informatikai tudományok | hu |
| dc.subject.sciencefield | Műszaki tudományok | hu |
| dc.title | Ensemble Methods in Medical Decision Making | hu_HU |
| dc.title.translated | Összetett módszerek az orvosi döntéshozatalban | hu_HU |