QRS complex detection is performed on the first five minutes of all 48 electrocardiogram (ECG) signals of the MIT-BIH Arrhythmia database (17). A wavelet packet decomposition and compression algorithm optimized for a particular channel (MLII lead) in this database is used for detecting the QRS complex. On the MLII lead, the algorithm on average shows a sensitivity of 96.4% and a predictive value of a positive test of 98.1%. For three signals, other channels were analyzed and the average sensitivity was 92.5% and the average predictive value of a positive test was 95.8%.; To evaluate the hypothesis that the area to duration ratio of the QRS complex in normal heartbeat segments can be used as a predictor for arrhythmias, ECG's of healthy patients and the normal segment of ECG's of patients with arrhythmias were analyzed. It was found that the ratio does not seem to be a suitable predictor of abnormality for this data set. |