Please use this identifier to cite or link to this item:
|Title:||Wrist Pulse Signal Processing for Inflammation Diagnostic Analysis|
|Authors:||Chow, Wai Hei|
|Department:||Department of Electronic Engineering|
|Supervisor:||Supervisor: Dr. TSANG, K F; Assessor: Prof. CHAN, Y C|
|Abstract:||Palpation, which is the process of using the fingers to examine the wrist pulse, is the major diagnostic tool in Traditional Chinese Medicine (TCM). The wrist pulse is believed to contain critical information of the patients’ health condition. This project aims to analyze the time series wrist pulse signals in order to distinguish patients suffering from various symptoms with healthy people. In this project, inflammation is taken into account as the diagnosis of inflammation is still ineffective and inaccurate. The four inflammation symptoms tackled in thiis project are Appendicitis, Acute Appendicitis, Pancreatitis and Duodenal Bulb Ulcer. Moreover, studying the characteristic of blood flow in arteries and cardiac cycle is crucial for the sake of selecting features from the time series wrist pulse signals. First of all, the wrist pulse is captured by a Doppler Ultrasound sensor, which is to measure the blood flow velocity inside a blood vessel. Subsequently, features are extracted from the wrist pulse signals. The Doppler parameters and the regression coefficients obtained from Auto-regression (AR) model are defined as the disease sensitive features. In addition, an innovative idea is proposed in this project, which is to observe the characteristic of the blood flow acceleration time graph. Consequently, the features extracted are the parameter for training the Support Vector Machine (SVM) classifier in order to be able to classify the symptoms from healthy condition. The classification accuracy can reach over 90% in distinguishing patients with healthy persons from Acute Appendicitis and up to 99% from Appendicitis and Pancreatitis respectively. Furthermore, the accuracy up to 85.4% is obtained when distinguishing all the symptoms taken into account in addition to healthy persons simultaneously. These results indicate the methodology proposed in this project can provide an advanced idea for enhancing the research of wrist pulse signal analysis.|
|Appears in Collections:||Electronic Engineering - Undergraduate Final Year Projects |
Items in Digital CityU Collections are protected by copyright, with all rights reserved, unless otherwise indicated.