Ischemic stroke detection system with computer aided diagnostic capability
Loading...
Date
2017-06
Authors
Lina Tay
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Ischemic stroke is caused by narrowing of the blood vessel due to emboli travel
along the blood vessel that eventually trapped near the vessel wall and become stenosis.
Transcranial Doppler (TCD) Ultrasound is used as a tool to detect emboli. However, the
TCD monitoring process is time-consuming and fatigue. Since the evaluation requires
human experts, limited number of experts makes the manual emboli detection a
challenging task. Therefore, this project is to develop program for automated emboli
detection. MATLAB are used to develop signal processing algorithm of the system. In
this project, there are four detection methods investigated. The first method is
sinusoidal modelling method where the frequency spectrum were inspected to search
for the frequency components with high magnitude. The second method compares the
energy and zero crossing rate of embolic signal with the threshold level. Subsequently,
the short time energy and short time average zero crossing rate method is employed to
compare two characteristic with threshold level computed. The last method is the
Support Vector Machine (SVM) classifier where Mel Frequency Cepstral Coefficients
(MFCC) is the extracted features used to train the classifier. The performance
evaluations of the detection methods are measured by accuracy percentage and
processing time. The best result is achieved by the sinusoidal modelling method with
high genuine acceptance rate at 84.2% and low false rejection rate of 33.14%. After the
proposed software system is validated, the system is modified and employed into a
graphical user interface (GUI) application.