A wearable device-based ECG identity authentication method

Abstract

The invention discloses an electrocardiogram identity authentication method based on a wearable device. The method preprocesses the electrocardiogram input by the user, detects and extracts feature points, and matches and compares the extracted electrocardiogram features with the feature templates in the user database. Aiming at the problem of electrocardiogram noise reduction on wearable devices, the present invention designs a hybrid three-layer noise reduction that can effectively remove noise signals. The hybrid three-layer noise reduction combines the advantages of morphological filtering and frequency domain filtering. The signal quality of the obtained ECG can be significantly guaranteed. The user feature template constructed by the present invention retains the respective hard index and soft index characteristics of the reference feature and the non-reference feature of the electrocardiogram, thereby ensuring the authentication accuracy of the system.

Description

technical field [0001] The invention belongs to the field of biological authentication, and in particular relates to an electrocardiogram identity authentication method based on a wearable device. Background technique [0002] In recent years, wearable devices (such as smart watches) have been widely used in mobile payment, mobile communication, health information monitoring, smart home interaction and other fields. There are already smart watches with ECG measurement functions, such as Apple Watch and Huami Technology's health watch. [0003] In the past, the research data in ECG research were collected by ECG machines with high anti-interference and high precision. This type of ECG data has a high signal-to-noise ratio and no high-energy low-frequency noise and high-frequency noise. Therefore, previous studies on ECG recognition and authentication can remove noise in ECG with only one bandpass filter. However, this frequency-domain filtering method is not suitable for re...

AI Technical Summary

Problems solved by technology

However, this frequency-domain filtering method is not suitable for removing noise in ECG collected by wearable devices

The signal-to-noise ratio of the ECG collected by wearable devices is low, and the high-energy noise frequency band covers the key low-frequency area of ​​the ECG. Frequency domain filtering cannot effectively extract the frequency band representing the ECG signal from the frequency band where the signal and noise are mixed.

On the other hand, when constructing user feature templates, previous studies often ignore the characteristics of ECG features and directly construct feature templates. Few studies distinguish these two types of features, let alone adjust the two types of features in the user feature template. proportion, the feature template constructed in this way cannot fully represent the user's ECG information

Images