Traditional one-time user authentication processes might cause friction and unfavorable user experience in many widely-used applications. This is a severe problem in particular for security-sensitive facilities if an adversary could obtain unauthorized privileges after a user’s initial login. Recently, continuous user authentication (CA) has shown its great potential by enabling seamless user authentication with few active participation. We devise a low-cost system exploiting a user’s pulsatile signals from the photoplethysmography (PPG) sensor in commercial wrist-worn wearables for CA. Compared to existing approaches, our system requires zero user effort and is applicable to practical scenarios with non-clinical PPG measurements having motion artifacts (MA). We explore the uniqueness of the human cardiac system and design an MA filtering method to mitigate the impacts of daily activities. Furthermore, we identify general fiducial features and develop an adaptive classifier using the gradient boosting tree (GBT) method. As a result, our system can authenticate users continuously based on their cardiac characteristics so little training effort is required. Experiments with our wrist-worn PPG sensing platform on 20 participants under practical scenarios demonstrate that our system can achieve a high CA accuracy of over 90% and a low false detection rate of 4% in detecting random attacks.
more »
« less
This content will become publicly available on July 1, 2025
Inferring Electrocardiography From Optical Sensing Using Lightweight Neural Network
This article presents a computational solution that enables continuous cardiac monitoring through cross-modality inference of electrocardiogram (ECG). While some smartwatches now allow users to obtain a 30-s ECG test by tapping a built-in bio-sensor, these short-term ECG tests often miss intermittent and asymptomatic abnormalities of cardiac functions. It is also infeasible to expect persistently active user participation for long-term continuous cardiac monitoring in order to capture these and other types of cardiac abnormalities. To alleviate the need for continuous user attention and active participation, we design a lightweight neural network that infers ECG from the photoplethysmogram (PPG) signal sensed at the skin surface by a wearable optical sensor. We also develop a diagnosis-oriented training strategy to enable the neural network to capture the pathological features of ECG, aiming to increase the utility of reconstructed ECG signals for screening cardiovascular diseases (CVDs). We also leverage model interpretation to obtain insights from data-driven models, for example, to reveal some associations between CVDs and ECG/PPG and to demonstrate how the neural network copes with motion artifacts in the ambulatory application. The experimental results on three datasets demonstrate the feasibility of inferring ECG from PPG, achieving a high fidelity of ECG reconstruction with only about 40000 parameters.
more »
« less
- Award ID(s):
- 2124291
- PAR ID:
- 10538192
- Publisher / Repository:
- IEEE Transactions on Artificial Intelligence
- Date Published:
- Journal Name:
- IEEE Transactions on Artificial Intelligence
- Volume:
- 5
- Issue:
- 7
- ISSN:
- 2691-4581
- Page Range / eLocation ID:
- 3535 to 3550
- Subject(s) / Keyword(s):
- Cross-modality inference, electrocardiogram (ECG), neural network, photoplethysmogram (PPG), physiological digital twin, tele-health.
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Physiological and kinematic signals from humans are often used for monitoring health. Several processes of interest (e.g., cardiac and respiratory processes, and locomotion) demonstrate periodicity. Training models for inference on these signals (e.g., detection of anomalies, and extraction of biomarkers) require large amounts of data to capture their variability, which are not readily available. This hinders the performance of complex inference models. In this work, we introduce a methodology for improving inference on such signals by incorporating phase-based interpretability and other inference tasks into a multi-task framework applied to a generative model. For this purpose, we utilize phase information as a regularization term and as an input to the model and introduce an interpretable unit in a neural network, which imposes an interpretable structure on the model. This imposition helps us in the smooth generation of periodic signals that can aid in data augmentation tasks. We demonstrate the impact of our framework on improving the overall inference performance on ECG signals and inertial signals from gait locomotion.more » « less
-
Construction workers often experience high levels of physical and mental stress due to the demanding nature of their work on construction sites. Real-time health monitoring can provide an effective means of detecting these stressors. Previous research in this field has demonstrated the potential of photoplethysmography (PPG), which represents cardiac activities, as a biomarker for assessing various stressors, including physical fatigue, mental stress, and heat stress. However, PPG acquisition during construction tasks is subject to several external noises, of which motion artifact is a major one. To address this, the study develops and examines an autoencoder network—a special type of artificial neural network—to remove PPG signals’ motion artifacts during construction tasks, thereby enhancing the accuracy of health assessments.Artifact-free PPG signals are acquired through subjects in a stationary position, which is used as the reference for training the autoencoder network. The network’s performance is examined with PPG signals acquired from the same subjects performing multiple construction tasks. The developed autoencoder network can increase the signal-to-noise ratio (SNR) by up to 33% for the corrupted signals acquired in a construction setting. This research contributes to the extensive and resilient use of PPG signals in health monitoring for construction workers.more » « less
-
more » « less
Abstract Heart rhythm assessment is indispensable in diagnosis and management of many cardiac conditions and to study heart rate variability in healthy individuals. We present a proof-of-concept system for acquiring individual heart beats using smart speakers in a fully contact-free manner. Our algorithms transform the smart speaker into a short-range active sonar system and measure heart rate and inter-beat intervals (R-R intervals) for both regular and irregular rhythms. The smart speaker emits inaudible 18–22 kHz sound and receives echoes reflected from the human body that encode sub-mm displacements due to heart beats. We conducted a clinical study with both healthy participants and hospitalized cardiac patients with diverse structural and arrhythmic cardiac abnormalities including atrial fibrillation, flutter and congestive heart failure. Compared to electrocardiogram (ECG) data, our system computed R-R intervals for healthy participants with a median error of 28 ms over 12,280 heart beats and a correlation coefficient of 0.929. For hospitalized cardiac patients, the median error was 30 ms over 5639 heart beats with a correlation coefficient of 0.901. The increasing adoption of smart speakers in hospitals and homes may provide a means to realize the potential of our non-contact cardiac rhythm monitoring system for monitoring of contagious or quarantined patients, skin sensitive patients and in telemedicine settings.
-
An Empirical Study Comparing Unobtrusive Physiological Sensors for Stress Detection in Computer WorkSeveral unobtrusive sensors have been tested in studies to capture physiological reactions to stress in workplace settings. Lab studies tend to focus on assessing sensors during a specific computer task, while in situ studies tend to offer a generalized view of sensors’ efficacy for workplace stress monitoring, without discriminating different tasks. Given the variation in workplace computer activities, this study investigates the efficacy of unobtrusive sensors for stress measurement across a variety of tasks. We present a comparison of five physiological measurements obtained in a lab experiment, where participants completed six different computer tasks, while we measured their stress levels using a chest-band (ECG, respiration), a wristband (PPG and EDA), and an emerging thermal imaging method (perinasal perspiration). We found that thermal imaging can detect increased stress for most participants across all tasks, while wrist and chest sensors were less generalizable across tasks and participants. We summarize the costs and benefits of each sensor stream, and show how some computer use scenarios present usability and reliability challenges for stress monitoring with certain physiological sensors. We provide recommendations for researchers and system builders for measuring stress with physiological sensors during workplace computer use.more » « less
