Discrete Wavelet-based Fuzzy Network Architecture for ECG Rhythm-Type Recognition: Feature Extraction and Clustering- Oriented Tuning of Fuzzy Inference System

The paper addresses a new QRS complex geometrical feature extraction technique as well as its application for supervised electrocardiogram (ECG) heart-beat type classification. Toward this objective, after detection and delineation of major events of the ECG signal via an appropriate algorithm, each QRS region and also its corresponding discrete wavelet transform (DWT) are supposed as virtual images and each of them is divided into eight polar sectors. Then, the curve length of each excerpted segment is calculated and is used as the element of the feature space. Afterwards, an appropriate fuzzy network classifier aimed for recognizing several heart-beat types is preliminarily designed. To propose a new classification strategy with adequate robustness against noise, artifacts and arrhythmic outliers, the fuzzy rules parameterization and determination stages were fulfilled using the fuzzy c-means (FCM) and the subtractive clustering techniques. To show merit of the new proposed algorithm, it was applied to 4 number of arrhythmias (Normal, Left Bundle Branch Block-LBBB, Right Bundle Branch Block-RBBB, Paced Beat-PB) belonging to 12 records of the MIT-BIH Arrhythmia Database and the average accuracy values Acc=94.58% and Acc=97.41% were achieved for FCM-based and subtractive clustering-based fuzzy-logic classifiers, respectively. To evaluate operating characteristics of the new proposed fuzzy classifier, the obtained results were compared with similar peer-reviewed studies in this area.