ECG features and methods for automatic classification of ventricular premature and ischemic heartbeats: A comprehensive experimental study

Abstract

Accurate detection of cardiac pathological events is an important part of electrocardiogram (ECG) evaluation and subsequent correct treatment of the patient. The paper introduces the results of a complex study, where various aspects of automatic classification of various heartbeat types have been addressed. Particularly, non-ischemic, ischemic (of two different grades) and subsequent ventricular premature beats were classified in this combination for the first time. ECGs recorded in rabbit isolated hearts under non-ischemic and ischemic conditions were used for analysis. Various morphological and spectral features (both commonly used and newly proposed) as well as classification models were tested on the same data set. It was found that: a) morphological features are generally more suitable than spectral ones; b) successful results (accuracy up to 98.3% and 96.2% for morphological and spectral features, respectively) can be achieved using features calculated without time-consuming delineation of QRS-T segment; c) use of reduced number of features (3 to 14 features) for model training allows achieving similar or even better performance as compared to the whole feature sets (10 to 29 features); d) k-nearest neighbours and support vector machine seem to be the most appropriate models (accuracy up to 98.6% and 93.5%, respectively).

SVM Based ECG Classification Using Rhythm and Morphology Features, Cluster Analysis and Multilevel Noise Estimation

“…SVM is widely used for ECG classification due to its simplicity, robustness and efficiency [2,3], which was confirmed in our previous study, too [4,5].…”

Advanced P Wave Detection in Ecg Signals During Pathology: Evaluation in Different Arrhythmia Contexts

“…These features are: a) the area under the curve (AUC) – calculated in the segment of the original ECG demarcated from R ( i )−50 ms to R ( i )+50 ms, where R ( i ) is the position of the current R wave; b) the difference between AUC of the actual and the previous QRS complex; c) the length of RR ( i ); d) the length of RR ( i- 1); d) the difference between RR ( i ) and RR ( i- 1); e) the altitude of the actual QRS complex U QRS ( i ); f) the difference between the altitude of the actual and the previous QRS complex, U QRS ( i )− U QRS ( i- 1). Some of these features were previously published in studies 34,40 , where their usability was proven. All of these features can be computed in a simple manner, and they indicate significant differences between the normal heartbeats and PVC, as shown by the results of a nonparametric Kruskal-Wallis test (α = 0.05), followed by a Tukey-Kramer post hoc test 40 .…”

“…Some of these features were previously published in studies 34,40 , where their usability was proven. All of these features can be computed in a simple manner, and they indicate significant differences between the normal heartbeats and PVC, as shown by the results of a nonparametric Kruskal-Wallis test (α = 0.05), followed by a Tukey-Kramer post hoc test 40 . These features were further used as input in the classification algorithm, to distinguish between PVC and other beats.…”

“…For differentiation between PVC and other beats, the k-nearest neighbors (k-NN) method was used 40 . The k-NN algorithm is one of the simplest, and most-used machine learning classification algorithms.…”

Computational Diagnostic Techniques for Electrocardiogram Signal Analysis

“…Compared with the time-frequency transform method, appropriate structural elements can preserve the signal well. However, most of the statistical features and morphological features are not comprehensive and prominent, and most of the studies do not use these features alone, but often use them in combination with other kinds of features, or use statistical methods to further extract the existing features [77,[88][89][90]. Raj et al combined dictionary decomposition method with statistical characteristics, and proposed an overcomplete Gabor dictionary-based statistical feature extraction method [91].…”