Graduation Semester and Year

2005

Language

English

Document Type

Thesis

Degree Name

Master of Science in Computer Science

Department

Computer Science and Engineering

First Advisor

JungHwan Oh

Abstract

Capsule Endoscopy (CE) is a new procedure where endoscopists can visualize and discriminate anomalies of the human gastrointestinal track by allowing the patient to swallow a camera pill. The digital video obtained from the pill is used to safely pinpoint the location of these abnormalities in areas previously unavailable for view. One of the main functions of CE is determining the location of bleeding. Given Imaging, currently the only maker of a FDA approved CE device, has a proprietary and unspecified blood test, called Suspected Blood Index, which can be used to automatically detect where blood occurs in the video. Its precision and recall are fair, but can be improved upon. In this paper we discuss the journey taken in the attempt to attempt to create a better blood detector. The final algorithm seems to deliver good results with precision and recall above 92% and 85% respectively.

Disciplines

Computer Sciences | Physical Sciences and Mathematics

Comments

Degree granted by The University of Texas at Arlington

Download

Share

COinS