Detection method for shockable rhythm based on a single feature

Abstract

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

Keywords

• shockable rhythm
• single feature extraction
• discrete wavelet transform
• single‐channel electrocardiogram
• support vector machine
• five‐fold cross validation