Use this URL to cite or link to this record in EThOS:
Title: Computer assisted detection and modelling of paediatric airway pathology from medical images
Author: Irving, B. J.
Awarding Body: University College London (University of London)
Current Institution: University College London (University of London)
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
This thesis proposes a novel computer assisted detection framework to analyse airway shape change and to detect signs of paediatric pulmonary tuberculosis (TB). The method can accurately distinguish TB patients with airway involvement, from non-TB patients using CT scans. This model is also applied to X-ray radiographs to segment the airways. As a first step, a CT airway segmentation algorithm is proposed, and then evaluated as part of the EXACT’09 airway segmentation challenge. Algorithms are then implemented to extract the airway centreline, detect branch points and label each airway branch. A number of cases had the appearance of complete obstruction in some bronchi, and a method is presented to identify and segment beyond these obstructions. A statistical shape model is developed using airway shape variation, which requires correspondence between airways. Thus, a method to register regions of the airway tree is proposed. This method generates landmarks on the airway surface. Using these landmarks, a template mesh is then aligned to each airway by thin-plate-spline warp and local vertex alignment. This develops a corresponding surface mesh representation for each airway that can be used for statistical analysis. A statistical model of the variation of local regions of the airway is constructed and features derived from this are used to train a classifier to detect abnormal airway variation. The method is able to accurately detect TB in unseen cases. It is also compared to second method based on automated bronchi measurements. Finally, a method is developed that uses the previous 3D model to segment the airways in 2D radiographs. This uses an optimisation algorithm to fit the 3D model to image features in the radiograph. These methods create novel tools for airway analysis and have wider applications in medical imaging.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available