A Novel Abnormal Segments Detection Method for Photopletysmography Signal

There are many abnormal segments in photopletysmography (PPG) signal which is recorded by wearable device, these abnormal segments still exist after filtering and have a bad impact on the accuracy of signal processing. Therefore, we proposed a novel method named abnormal segments classification detection (ASCD) to detect these abnomal segments. First, the abnormal segments is classified as pulse interference, missing segment and motion artifact. Then, different approches are proposed to detect these abnormal segments. The PPG signals recorded by wearable device are as the experimental data. The proposed method and two common used methods are employed to detect the abnormal segments from the experimental data. The results show that the proposed method with the accuracy of 95.4±1.4%, the sensitivity of 97.2±0.8% and the specificity of 88.6±1.3%, which are more higher than that of commen methods. Therefore, the proposed method can more effectivly detect the abnormal segments from PPG signal and is potentail for wearable device applications.