A Computer-aided Diagnostic System for Wireless Capsule Endoscopy
Loading...
Authors
Chen, Ying-ju
Issue Date
2012-11
Type
Thesis
Language
en_US
Keywords
Engineering , Biomedical engineering , Medical imaging , Computer science , Applied sciences , Computer aided diagnosis , Wireless capsule endoscopy , Ulcer detection , Energy based event boundary detection , Bleeding detection , Health sciences , Environmental science
Alternative Title
Abstract
Wireless capsule endoscopy (WCE) is a technology breakthrough that allows the non-invasive visualization of the entire small intestine. This swallowable capsule technology enables the investigation of the small intestine without pain or need for sedation, thus encouraging patients to undergo gastrointestinal (GI) track examinations. Nevertheless, the average viewing time ranges from one to two hours, depending on the experience of the gastroenterologist. To address this problem, we look into machine vision based approaches for automatic medical imaging analysis. The process of selecting suitable low-level features is often determined by the modality of imaging as well as the nature of diagnoses. For example, image arises from radiologic projection technique such as x-ray often shows overlapping structures while image from tomographic technique such as magnetic resonance imaging (MRI) shows individually bounded anatomic objects. Computational approaches for projection based images may concentrates on detecting density blobs while tomography based images may focus on the identifying the relationships among different organ boundaries. Each imaging modality imposes unique restrictions for automatic image abstraction processes. In this dissertation, we investigate low-level image features for automatic analysis in WCE imaging, which is photography based imaging modality, and propose a computer-aided diagnostic (CAD) system for indexing and assisting physicians to speed up the diagnosis time. The advancing in Very-large-scale integration (VLSI), broadband networks, and image/video standards dramatically increases the amount of digital library and multimedia database. However, a robust image understanding at the machine level remains an open issue. Today, most of the image search engines perform keyword based textual search on text annotation for visual information. Due to lacking of automatic annotation system, today's search approach on medical imaging still relies on manual annotation and hence, it is not scalable. Manual annotation is not only laborious, error prone, very costly, if a collection of images are not annotated, they are not indexed or retrieved. The aims of this system are to help physicians reducing diagnosis time with an automatic tool that detects medical significant events in the WCE domain. In addition, this system is also served as a critical piece to bridging the semantic gap between machine and human perception in medical imaging.
Description
This thesis is being archived as a Digitized Shelf Copy for campus access to current students and staff only. We currently cannot provide this open access without the author's permission. If you are the author of this work and desire to provide it open access or wish access removed please contact the Wahlstrom Library to discuss permission.
Citation
Y. Chen, "A Computer-aided Diagnostic System for Wireless Capsule Endoscopy", Ph.D. dissertation, Dept. of Computer Science and Engineering, Univ. of Bridgeport, Bridgeport, CT, 2012.