作者: Jianchu Yao , Yongbo Wan
DOI:
关键词:
摘要: This paper presents a neural network developed to identify human subjects using electrocardiogram (ECG) signals collected from an "in-house" wearable sensor. In this project, noises were first removed the raw with wavelet filters. ECG cycles then extracted filtered and decomposed into coefficient structures. These structures used as input vectors 3-layer feedforward that generates identification results. current study, 61 datasets 23 utilized train network, which thereafter was tested 15 new different subjects. All in experiment successfully identified. The testing results demonstrate is effective. Keywords—Coefficient structure, electrocardiogram, training, decomposition