Cardiovascular Signal Processing
Speaker: Prof. Samaren Dandapat, Dept. of ECE, IIT Guwahati
The human cardiovascular system consists of the heart and blood vessels which facilitate the circulation of blood throughout the body. Bioelectric signals are responsible for mechanical events of the heart. This tutorial will focus on three topics. First, the electrocardiogram (ECG) and the phonocardiogram (PCG) signals will be discussed with emphasis on their physiological correlations. The ECG signal represents the changes in the electrical potential in the cardiac muscle. Each ECG cycle contains a number of time-varying local waves which provide important information on different cardiac events. The PCG is a recording of the acoustic waves produced by the mechanical action of the heart. It is a very useful and effective method for the diagnosis of heart valves.
Secondly, we will present a discussion on different quality measures used for the evaluation of cardiovascular signals. Processing of signals such as compression, transmission etc. introduces distortions in the signals. It is important to preserve the clinical information in the signal during such processing. Some of the diagnostic distortion measures namely the weighted percentage root mean square difference (WPRD), weighted diagnostic distortion (WDD) and wavelet weighted PRD (WWPRD) will be presented.
Thirdly, we will discuss some research directions in the area of ECG and PCG compression. The presentation will cover the discrete wavelet transform based cardiovascular signal compression. A major problem in the evaluation of lossy methods is the difficulty in assessing the clinical distortions using traditional objective distortion measures. Efficient compression algorithms can be designed using recently proposed diagnostic distortion measures.