ECG in real world scenario: time variability in biometric using wearable smart textile shirts
Biomedical signals, such as an electrocardiogram (ECG), have been included in wearable platforms for biometric reasons due to the rapid expansion of apps and technology capable of gathering this physiological data. Most studies using ECG biometrics are conducted in clinical settings, which is imp...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English English |
| Published: |
Semarak Ilmu Publishing
2024
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/111227/2/111227_ECG%20in%20real%20world%20scenario.pdf http://irep.iium.edu.my/111227/8/111227_ECG%20in%20real%20world%20scenario_SCOPUS.pdf http://irep.iium.edu.my/111227/ https://semarakilmu.com.my/journals/index.php/applied_sciences_eng_tech/issue/view/411 https://doi.org/10.37934/araset.40.2.3649 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Islam Antarabangsa Malaysia |
| Language: | English English |
| Summary: | Biomedical signals, such as an electrocardiogram (ECG), have been included in wearable
platforms for biometric reasons due to the rapid expansion of apps and technology
capable of gathering this physiological data. Most studies using ECG biometrics are
conducted in clinical settings, which is impractical for wearable ECG-based biometric
applications. Therefore, this study aims to determine the reliability of ECG signals
obtained from the commercially available Hexoskin Proshirt and HeartIn Fit shirt, which
may be worn for biometric verification in real-world scenarios. ECG data from 22
participants were collected over a span period of more than 30 days. The raw ECG signal
is first pre-processed in the time domain using noise-removal Butterworth filters, and
then a successful QRS segmented feature extraction method is used. Not to mention,
300 datasets were used to test the recommended recognition method using a Quadratic
Support Vector Machine (QSVM). In comparison, around 854 datasets were prepared
for training and validation of the classifier. The findings showed that the proposed
method provided a considerable accuracy above 99.63 % with a FAR of 0.14 %, an FRR
of 2.86 %, and a TPR of 97.14 %. Thus, the study supports using ECG biometrics for
verification in real-world settings by employing a smart textile shirt with varying
temporal variability. |
|---|