Biometrie identification from raw ECG signal using deep learning techniques

Yuriy Khoma, Dmytro Sabodashko, Veronika Herasymenko

2017 · 46 citations · 6 references

Concepts

Abstract

The paper presents and discusses a novel method of biometrie identification based on ECG data. The main idea of the study is to apply Deep Neural Networks (DNN) for human identification based on the raw ECG signal. To improve overall system accuracy various signal pre-processing and outlier detection techniques have been applied. Also, to make ECG identification approach more user friendly, three-finger measurement scheme has been proposed. All experiments have been made using self-collected Lviv Biometric Data Set.

References

6