Digital methods for ECG analysis

Abstract

The objective of the present thesis is to develop digital processing techniques for ECG signal analysis to be implemented on a dedicated system in a clinical environment. A computationally efficient algorithm for event detection and rhythm monitoring, which can be implemented on a microprocessor system, is developed. Also, a linear predictor model for QRS complexes of ECG signals is proposed. The model parameters, viz. linear predictor coefficients (LPCs), are made use of for classification of these complexes into normal and abnormal categories.

Chapter 1 presents a brief survey of the state-of-the-art of ECG signal processing and patient monitoring in intensive coronary care units using digital computers.

A general description of the data acquisition system used in the present work and the characteristics of various subsystems is given in Chapter 2.

The development of an algorithm for event detection and rhythm monitoring in ECG signals is considered in Chapter 3. For this purpose, a pair of transforms for detection of Q and S points of the QRS complex and start and end points of both P and T waves is proposed. A set of recursive relations is developed for making the transforms computationally efficient.