Life-Threatening Ventricular Arrhythmia Detection with Personalized Features

Date

2017

Authors

Cheng, Ping
Dong, Xiaodai

Journal Title

Journal ISSN

Volume Title

Publisher

IEEE Access

Abstract

The timely detection of life-threatening ventricular arrhythmias (VAs) is critical for saving a patient's life. General features that characterize ECG waveforms are extracted for VA detection. To take into account the subtle differences in the QRS complexes among different people, new personalized features are proposed in this paper based on the correlation coefficient between a patient-specific regular QRS-complex template and his/her real-time ECG data. Small sets of the most effective features are chosen with support vector machines from 11 newly extracted and 15 previously existing features, for efficient performance and real-time operation. Our proposed new features aveCC and medianCC are verified to be effective in enhancing the performance of existing features under both the record-based and database-based data divisions. Through 50-time random record-based data divisions, all combinations of two features and three features are tested. The top two-feature combination is VFleak and aveCC, which achieves an area under curve (AUC) value of 98.56%+/- 0.89%, a specificity (SP) of 94.80%+/- 2.15%, and an accuracy (ACC) of 94.66%+/- 1.97%; the top three-feature combination is VFleak, MEA, and aveCC, which obtains an AUC of 98.98%+/- 0.58%, an SP of 95.56%+/- 1.45%, and an ACC of 95.46%+/- 1.36%; these results outperform the previous top-two and top-three feature combinations. Similar results are obtained on the database-based data division.

Description

Keywords

correlation coefficient, ventricular arrhythmia detection, ECG, feature selection, personalized feature, patient-specific template matching, real-time classification, machine learning

Citation

Cheng, P. & Dong, X. (2017). Life-threatening ventricular arrhythmia detection with personalized features. IEEE Access, 5, 14195-14203. https://doi.org/10.1109/ACCESS.2017.2723258