Object-based Image Analysis for Detection and Segmentation Tasks in Biomedical Imaging

  • Object-based image analysis (OBIA) is a concept for analyzing images based on regions instead of pixels. OBIA allows to effectively incorporate features of regions as well as their contextual and hierarchical relations into the analysis process. While object-based image analysis is common in the field of geographic information science and remote sensing, it has rarely found its was into biomedical image analysis. This thesis explores the applicability and capabilities of OBIA for addressing image processing and analysis tasks on biomedical images, by approaching several relevant and challenging tasks from different imaging domains. First a formalization as well as a powerful and flexible implementation of the OBIA concept is proposed. This is the foundation on which the applications are based. The addressed applications are: detection of the spine and vertebrae in CT images; detection of pregnancy in pigs on ultrasound images; reconstruction of vessels from histological whole slide sections of murine liver samples. One of the major contributions of this thesis is the demonstration of the capability and applicability of OBIA to tackle biomedical image analysis problems. Together with the OBIA foundation, the presented solutions for three different applications reveal the strength of OBIA, but also some challenges. Furthermore, the developed algorithms also pose a valuable scientific contribution in their own right, with some of them even presenting world-novel algorithms that have already found their way into commercial application.

Download full text

Cite this publication

  • Export Bibtex
  • Export RIS

Citable URL (?):

Search for this publication

Search Google Scholar Search Catalog of German National Library Search OCLC WorldCat Search Bielefeld Academic Search Engine
Meta data
Publishing Institution:IRC-Library, Information Resource Center der Jacobs University Bremen
Granting Institution:Jacobs Univ.
Author:Michael Schwier
Referee:Horst Karl Hahn, Herbert Jaeger, Gitta Domik-Kienegger
Advisor:Horst Karl Hahn
Persistent Identifier (URN):urn:nbn:de:gbv:579-opus-1006017
Document Type:PhD Thesis
Date of Successful Oral Defense:2016/07/14
Date of First Publication:2016/08/18
Academic Department:Computer Science & Electrical Engineering
PhD Degree:Computer Science
Focus Area:Mobility
Other Organisations Involved:Fraunhofer MEVIS
Library of Congress Classification:R Medicine / R Medicine (General) / R858-859.7 Computer applications to medicine. Medical informatics
Call No:Thesis 2016/25

$Rev: 13581 $