Multi-slice Nested Recurrence Plot (MsNRP): A robust approach for person identification using daily ECG or PPG signals
YeongJun Jeon, Soon Ju Kang
Abstract
This study presents a novel approach called Multi-slice Nested Recurrence Plot (MsNRP) for person identification using noisy bio-signals. Prior studies in biometrics have predominantly relied on ideal datasets of 5–10 min, which introduces uncertainty in accuracy when dealing with noisy bio-signals. The proposed MsNRP method captures features from one and multiple cycles without the need for preprocessing, making it well-suited for photoplethysmograms(PPG) and electrocardiograms(ECG). By overcoming the limitations of traditional recurrence plots(RP), MsNRP demonstrates robustness to noisy bio-signal datasets, thus offering a reliable solution for identification in practical scenarios. We demonstrate the experiments of MsNRP on both 5–10 min datasets, similar to previous related work and day-long datasets, measured in daily life emphasizing its robustness in handling noisy data.