Lista przedmiotów z materiałami udostępnionymi dla studentów

Dla_studentów
  • Increase font size
  • Default font size
  • Decrease font size

Marek Rudnicki


A real time DSP implementation of wavelet transform - based QRS complex detector


Opiekun pracy dyplomowej: prof. dr hab. inż. Paweł Strumiłło
Praca dyplomowa ife obroniona 2006-07-25
Streszczenie pracy dyplomowej:
The problem of cardiovascular system diseases and in particular heart disorders is becoming more and more significant for the population. The most common way of diagnosis of the heart's condition is analysis of electrocardiogram (ECG). Traditionally ECG records are analyzed by physicians, but thanks to rapid development in the computer technology, the design of automated ECG analysis systems is possible.. In this way, the long-term (e.g. 24 hour) ECG records can be analyzed reliably within seconds. Another application for such systems is on-line monitoring of patients in unstable condition in the Intensive Care Units in hospitals. One of the most important tasks in the processing of ECG signals is the detection of heart-beats. In ECG records they correspond to characteristic pulse-like waveforms called QRS complexes. QRS complex detection is a crucial step in the determination of the heart rhythm and is necessary for further ECG signal processing in many cases. This thesis presents work on the development of an on-line QRS complex detector. The proposed algorithm is based on the discrete wavelet transform (DWT) and contains adaptive mechanisms that make it robust to many kinds of noise and to varying morphology of QRS complexes. Additionally, it is implemented on the Texas Instruments TMS320C6713 DSP development board. The testing procedure includes both the performance and the computational load measurements. The designed QRS complex detector is tested against a standard ECG database --- MIT-BIH Arrhythmia Database. The designed system has sensitivity of 99.42% and positive predictivity equal to 99.40% for the MIT-BIH database. This is a promising result and it is comparable with other available heart-beat detectors. Additional advantage is the low computational load of the DWT used in the processing.