Detection method for shockable rhythm based on a single feature

• DOI: 10.1049/el.2016.0344
• Publication Date: 2016-02-23T15:29:12Z
• Authors: D.H. Nam; D.W. Kang; H.S. Myoung; K.J. Lee
• Keywords: Feature (linguistics)

A method for detecting shockable rhythm was proposed. single feature extracted using the discrete wavelet transform from a single-channel electrocardiogram. The applied to support vector machine determine whether signal or non-shockable rhythm. To evaluate performance, five-fold cross validation performed with different window lengths (WLs). As result, sensitivity of 97.8% and specificity 98.0% 3 s WL were obtained. results represent proposed shows high detection performance even short analysis only feature.