Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.797533
Title: Computed tomography image analysis for the detection of obstructive lung diseases
Author: Chabat, Francois
Awarding Body: University of London
Current Institution: Imperial College London
Date of Award: 2000
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Damage to the small airways resulting from direct lung injury or associated with many systemic disorders is not easy to identify. Non-invasive techniques such as chest radiography or conventional tests of lung function often cannot reveal the pathology. On Computed Tomography (CT) images, the signs suggesting the presence of obstructive airways disease are subtle, and inter- and intra-observer variability can be considerable. The goal of this research was to implement a system for the automated analysis of CT data of the lungs. Its function is to help clinicians establish a confident assessment of specific obstructive airways diseases and increase the precision of investigation of structure/function relationships. To help resolve the ambiguities of the CT scans, the main objectives of our system were to provide a functional description of the raster images, extract semi-quantitative measurements of the extent of obstructive airways disease and propose a clinical diagnosis aid using a priori knowledge of CT image features of the diseased lungs. The diagnostic process presented in this thesis involves the extraction and analysis of multiple findings. Several novel low-level computer vision feature extractors and image processing algorithms were developed for extracting the extent of the hypo-attenuated areas, textural characterisation of the lung parenchyma, and morphological description of the bronchi. The fusion of the results of these extractors was achieved with a probabilistic network combining a priori knowledge of lung pathology. Creating a CT lung phantom allowed for the initial validation of the proposed methods. Performance of the techniques was then assessed with clinical trials involving other diagnostic tests and expert chest radiologists. The results of the proposed system for diagnostic decision-support demonstrated the feasibility and importance of information fusion in medical image interpretation.
Supervisor: Hansell, David M. ; Yang, Guang-Zhong Sponsor: Imatron, Inc.
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.797533  DOI: Not available
Share: