Search Results (158)

This study aimed to develop a non-invasive blood glucose estimation method by integrating wearable multimodal signals, including photoplethysmography (PPG), electrodermal activity (EDA), and skin temperature (ST), with food log–derived nutritional features, and to validate its clinical reliability. We analyzed data from 16 adults who underwent continuous glucose monitoring (CGM) while multimodal physiological signals were collected over 8–10 consecutive days, yielding more over 20,000 paired samples. Features from food logs and physiological signals were extracted, followed by feature selection using Boruta and minimum Redundancy Maximum Relevance (mRMR). Five machine learning models were trained and evaluated using five-fold cross-validation. Food log features alone demonstrated stronger predictive power than unimodal physiological signals. The fusion of nutritional, physiological, and temporal features achieved the best accuracy using LightGBM, reducing the RMSE to 12.9 mg/dL, with a MARD of 7.9%, a MAE of 8.82 mg/dL, and R² of 0.69. SHapley Additive exPlanations (SHAP) analysis revealed that 24-h carbohydrate and sugar intake, time since last meal, and short-term EDA features were the most influential predictors. By integrating multimodal wearable and dietary information, the proposed framework significantly enhances non-invasive glucose estimation. The interpretable LightGBM model demonstrates promising clinical utility for continuous monitoring and early dysglycemia management. Full article

Drowsy driving and sudden medical emergencies are major contributors to traffic accidents, necessitating continuous, non-intrusive driver monitoring. Since current technologies often struggle to balance accuracy with practicality, this study presents the design, fabrication, and validation of a smart steering wheel prototype. The device integrates dry-contact electrocardiogram (ECG), photoplethysmography (PPG), and inertial sensors to facilitate multimodal physiological monitoring. The system underwent a two-stage evaluation involving a single participant: laboratory validation benchmarking acquired signals against medical-grade equipment, followed by real-world testing in a custom electric research vehicle to assess performance under dynamic conditions. Laboratory results demonstrated that the prototype captured high-quality signals suitable for reliable heart rate variability analysis. Furthermore, on-road evaluation confirmed the system’s operational functionality; despite increased noise from motion artifacts, the ECG signal remained sufficiently robust for continuous R-peak detection. These findings confirm that the multimodal smart steering wheel is a feasible solution for unobtrusive driver monitoring. This integrated platform provides a solid foundation for developing sophisticated machine-learning algorithms to enhance road safety by predicting fatigue and detecting adverse health events. Full article

Pulsoximeters are widely used in the medical care of preclinical patients to evaluate the cardiorespiratory status and monitor basic vital signs, such as pulse rate (PR) and oxygen saturation (SpO₂). In many preclinical situations, air transport of the patient by helicopter is necessary. Conventional pulse oximeters, mostly used on the patient’s finger, are prone to motion artifacts during transportation. Therefore, this study aims to determine whether simulated helicopter vibration has an impact on the photoplethysmogram (PPG) derived from an in-ear sensor at the external ear canal and whether the vibration influences the calculation of vital signs PR and SpO₂. The in-ear PPG signals of 17 participants were measured at rest and under exposure to vibration generated by a helicopter simulator. Several signal quality indicators (SQI), including perfusion index, skewness, entropy, kurtosis, omega, quality index, and valid pulse detection, were extracted from the in-ear PPG recordings during rest and vibration. An intra-subject comparison was performed to evaluate signal quality changes under exposure to vibration. The analysis revealed no significant difference in any SQI between vibration and rest (all p > 0.05). Furthermore, the vital signs PR and SpO₂ calculated using the in-ear PPG signal were compared to reference measurements by a clinical monitoring system (ECG and SpO₂ finger sensor). The results for the PR showed substantial agreement (${\mathrm{CCC}}_{rest}$ = 0.96; ${\mathrm{CCC}}_{vibration}$ = 0.96) and poor agreement for SpO₂ (${\mathrm{CCC}}_{rest}$ = 0.41; ${\mathrm{CCC}}_{vibration}$ = 0.19). The results of our study indicate that simulated helicopter vibration had no significant impact on the calculation of the SQIs, and the calculation of vital signs PR and SpO₂ did not differ between rest and vibration conditions. Full article

This work presents a unified benchmarking framework for evaluating classical machine-learning–based heart-rate estimation from remote photoplethysmography (rPPG) across both RGB and near-infrared (NIR) modalities. Despite extensive research on algorithmic rPPG methods, their relative robustness across datasets, illumination conditions, and sensor types remains inconsistently reported. To address this gap, we standardize ROI extraction, signal preprocessing, rPPG computation, handcrafted feature generation, and label formation across four publicly available datasets: UBFC-rPPG Part 1, UBFC-rPPG Part 2, VicarPPG-2, and IMVIA-NIR. We benchmark five rPPG extraction methods (Green, POS, CHROM, PBV, PCA/ICA) combined with four classical regressors using MAE, RMSE, and $R^2$, complemented by permutation feature importance for interpretability. Results show that CHROM is consistently the most reliable algorithm across all RGB datasets, providing the lowest error and highest stability, particularly when paired with tree-based models. For NIR recordings, PCA with spatial patch decomposition substantially outperforms ICA, highlighting the importance of spatial redundancy when color cues are absent. While handcrafted features and classical regressors offer interpretable baselines, their generalization is limited by small-sample datasets and the absence of temporal modeling. The proposed pipeline establishes robust cross-dataset baselines and offers a standardized foundation for future deep-learning architectures, hybrid algorithmic–learned models, and multimodal sensor-fusion approaches in remote physiological monitoring. Full article

Remote photoplethysmography (rPPG) enables non-contact heart rate estimation but remains highly sensitive to illumination variation and dataset-dependent factors. This study proposes CHROM-Y, a robust 2D feature representation that combines chrominance (Ω, Φ) with luminance (Y) to improve physiological signal extraction under varying lighting conditions. The proposed features were evaluated using U-Net, ResNet-18, and VGG16 for heart rate estimation and waveform reconstruction on the UBFC-rPPG and BhRPPG datasets. On UBFC-rPPG, U-Net with CHROM-Y achieved the best performance with a Peak MAE of 3.62 bpm and RMSE of 6.67 bpm, while ablation experiments confirmed the importance of the Y-channel, showing degradation of up to 41.14% in MAE when removed. Although waveform reconstruction demonstrated low Pearson correlation, dominant frequency preservation enabled reliable frequency-based HR estimation. Cross-dataset evaluation revealed reduced generalization (MAE up to 13.33 bpm and RMSE up to 22.80 bpm), highlighting sensitivity to domain shifts. However, fine-tuning U-Net on BhRPPG produced consistent improvements across low, medium, and high illumination levels, with performance gains of 11.18–29.47% in MAE and 12.48–27.94% in RMSE, indicating improved adaptability to illumination variations. Full article

Remote photoplethysmography (rPPG) is a non-contact technology that estimates physiological signals, such as Heart Rate (HR), by capturing subtle skin color changes caused by periodic blood volume variations using only a standard RGB camera. While cost-effective and convenient, it suffers from a fundamental limitation: performance degrades severely in dynamic environments due to susceptibility to noise, such as abrupt illumination changes or motion blur. This study presents a deep learning framework that combines two structural modifications to ensure robustness in dynamic environments, specifically modeling movement noise and illumination change noise. The proposed framework structurally cancels global disturbances, such as illumination changes or global motion, through a dual-branch pipeline that encodes the face and background in parallel after Video Color Magnification (VCM) and then performs differencing. Subsequently, it utilizes a structure that injects a Temporal Shift Module (TSM) into the Spatio-Temporal Feature Extraction (SSFE) block to preserve long- and short-term temporal correlations and smooth noise, even amidst short and irregular movements. We measured MAE, RMSE, and correlation on the standard dataset UBFC-rPPG under four noise conditions: clean, illumination change noise, Movement Noise, Both Noise and the real-world in-vehicle dataset MR-NIRP (Stationary and Driving). Experimental results showed that the proposed method achieved consistent error reduction and correlation improvement compared to the VS-Net baseline in the illumination change noise-only and combined noise environments (UBFC-rPPG) and in the high-noise driving scenario (MR-NIRP). It maintained competitive performance in motion-only noise. Conversely, a modest performance disadvantage was observed under clean conditions (UBFC) and quasi-clean stationary conditions (MR-NIRP), interpreted as a design trade-off focused on global noise cancellation and temporal smoothing. Ablation studies demonstrated that the dual-branch pipeline is the primary contributor under illumination change noise, while TSM is the key contributor under movement noise, and that the combination of both elements achieves optimal robustness in the most complex scenarios. Full article

Photoplethysmography (PPG) is widely used in health monitoring, but its reliability is often compromised by artifacts, limiting accurate peripheral arterial oxygen saturation (SpO2) estimation. Moreover, physiological and demographic factors can substantially alter PPG waveform morphology. We propose a lightweight, real-time predictive modeling approach that adapts to subject-specific PPG signal dynamics to improve monitoring robustness under conditions prone to artifacts. A total of 459 min of dual-wavelength PPG signals, together with reference SpO2 values, were collected from 17 healthy volunteers (2 female, 15 male, mean age $27\pm 3$ years old) undergoing controlled desaturation in the 85–100% range after being instructed to remain still. Cardiac pulses were segmented and decomposed into AC and DC components, and the adequacy of several signal models, ranging from sums of Gaussians to Fourier series, and polynomial expansions of different orders, was evaluated. A space of representative signal features was built from the best-performing model, and used to generate machine learning-based predictions for each pulse using the preceding four clean pulses. Predicted pulses could be directly compared with their originals, enabling accurate error estimation without simulated data. The predicted signals closely matched the originals, achieving mean $R^2$ scores above 0.9, and an SpO2 estimation RMSE of 1.28%. In practical use, the same approach could be applied to overcome artifact-corrupted segments if combined with a signal quality assessment module. Therefore, this algorithm provides a promising pathway toward more reliable SpO2 monitoring in wearable systems, particularly under hypoxic conditions. Full article

Background/Objectives: Complex emotions in daily life often arise as mixtures of basic emotions, but most emotion-recognition systems still target a small set of discrete states and rely on contact-based sensing. This study aimed (1) to examine whether four compound emotions—Positive Surprise, Negative Surprise, Positive Sadness, and Negative Sadness—defined by valence direction within basic emotion categories can be differentiated using heart rate variability (HRV), and (2) to evaluate the feasibility of a camera-based contactless system (Deep Health Vision System, DHVS) by comparing it with a reference chest-strap device (Polar H10). Methods: Ten healthy adults viewed video clips designed to induce the four complex emotions. HRV was recorded simultaneously using Polar H10 and a webcam-based rPPG implementation of DHVS. Two-minute baseline and during-stimulus segments were extracted, and change rates of standard HRV indices were computed. After each stimulus, participants reported Valence, Arousal, Dominance, and proportional basic-emotion composition. Statistical analyses examined within-condition HRV changes, associations between HRV and self-reports, differences across emotion/valence conditions, and concordance between DHVS and Polar H10. Results: Self-reports confirmed distinct affective profiles for the four compound emotions. Positive and Negative Surprise were associated with heart rate reduction, while Positive Sadness showed reduced total power; Negative Sadness yielded heterogeneous but nonsignificant HRV changes. Specific HRV indices demonstrated condition-dependent correlations with Valence, Arousal, and Dominance. LF/HF changes were more sensitive to emotion category (Surprise vs. Sadness), whereas total power changes were more sensitive to valence (positive vs. negative). DHVS partially reproduced Polar H10 HRV patterns, with clearer concordance under positive-valence conditions. Conclusions: HRV captures distinct autonomic signatures of complex emotions defined by valence direction and shows meaningful links with subjective affective evaluations. LF/HF and total power provide complementary information on emotion category and valence-related autonomic reactivity, supporting indicator-specific modeling strategies. DHVS shows preliminary feasibility as a contactless HRV sensing platform for complex emotion recognition, warranting further validation with larger samples and more robust rPPG processing. Full article

Worldwide less than half of adults with hypertension are diagnosed and treated (only 42%), in addition one in five adults with hypertension (21%) has the condition under control. In the American continent, cardiovascular diseases (CVD) are the leading cause of death and high blood pressure (hypertension) is responsible for 50% of CVD deaths. Only a few countries show a population hypertension control rate of more than 50%. In this experimental study, we trained 15 regression-type machine learning algorithms, including traditional and ensemble methods to assess their effectiveness in estimating arterial pressure using noninvasive photoplethysmographic (PPG) signals extracted from 110 study subjects, to identify the risk of hypertension and its correlation with arteriosclerosis. We analyzed the performance of each algorithm using the metrics MSE, MAE, RMSE, and r². A 10-fold cross-validation showed that the best algorithms for hypertension risk identification were LR, KNN, SVR, RF, LR Baggin, KNNBagging, SVRBagging, and DTBagging. On the other hand, the best algorithms for arterioclesrosis risk identification were LR, KNN, SVR, RF, LR Bagging, and DTBagging. These results suggest that this research is promising and offers valuable information on the acquisition and processing of PPG signals. However, as this is an experimental study, the effectiveness of our model needs to be validated with a larger database. On the other hand, this model represents a support tool for healthcare specialists in the early detection of cardiovascular health, allowing people to self-manage their health and seek medical attention at an early stage. Full article

The real-time, precise monitoring of physiological signals such as intracranial pressure (ICP) and arterial blood pressure (BP) holds significant clinical importance. However, traditional methods like invasive ICP monitoring and invasive arterial blood pressure measurement present challenges including complex procedures, high infection risks, and difficulties in continuous measurement. Consequently, learning-based prediction utilizing observable signals (e.g., BP/pulse waves) has emerged as a crucial alternative approach. Existing models struggle to simultaneously capture multi-scale local features and long-range temporal dependencies, while their computational complexity remains prohibitively high for meeting real-time clinical demands. To address this, this paper proposes a physiological signal prediction method combining composite feature preprocessing with multiscale modeling. First, a seven-dimensional feature matrix is constructed based on physiological prior knowledge to enhance feature discriminative power and mitigate phase mismatch issues. Second, a network architecture CNN-LSTM-Attention (CBAnet), integrating multiscale convolutions, long short-term memory (LSTM), and attention mechanisms is designed to effectively capture both local waveform details and long-range temporal dependencies, thereby improving waveform prediction accuracy and temporal consistency. Experiments on GBIT-ABP, CHARIS, and our self-built PPG-HAF dataset show that CBAnet achieves competitive performance relative to bidirectional long short-term Memory (BiLSTM), convolutional neural network-long short-term memory network (CNN-LSTM), Transformer, and Wave-U-Net baselines across Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Coefficient of Determination (R²). This study provides a promising, efficient approach for non-invasive, continuous physiological parameter prediction. Full article

Reliable analysis of remote photoplethysmography (rPPG) signals depends on identifying physiologically plausible pulses. Traditional approaches rely on clustering self-similar pulses, which can discard valid variability. Automating pulse quality assessment could capture the true underlying morphology while preserving physiological variability. In this manuscript, individual rPPG pulses were manually labelled as plausible, borderline and implausible and used to train multilayer perceptron classifiers. Two independent datasets were used to ensure strict separation between training and test data: the Vision-MD dataset (4036 facial videos from 1270 participants) and a clinical laboratory dataset (235 videos from 58 participants). Vision-MD data were used for model development with an 80/20 training–validation split and 5-fold cross-validation, while the clinical dataset served exclusively as an independent test set. A three-class model was evaluated achieving $F_1$-scores of 0.92, 0.24 and 0.79 respectively. Recall was highest for plausible and implausible pulses but lower for borderline pulses. To test separability, three pairwise binary classifiers were trained, with ROC-AUC > 0.89 for all three category pairs. When combining borderline and implausible pulses into a single class, the binary classifier achieved an $F_1$-score of 0.93 for the plausible category. Finally, usability analysis showed that automated labelling identified more usable pulses per signal than the previously used agglomerative clustering method, while preserving physiological variability. Full article

Photoplethysmogram (PPG) signals are increasingly utilized in wearable and mobile healthcare applications due to their non-invasive nature and ease of use in measuring physiological parameters, such as heart rate, blood pressure, and oxygen saturation. Recent advancements have highlighted green-light photoplethysmogram (gPPG) as offering superior signal quality and accuracy compared to traditional red-light photoplethysmogram (rPPG). Given the deterministic chaotic nature of PPG signals’ dynamics, nonlinear time series analysis has emerged as a powerful method for extracting health-related information not captured by conventional linear techniques. However, optimal data conditions, including appropriate sampling frequency and minimum required time series length for effective nonlinear analysis, remain insufficiently investigated. This study examines the impact of downsampling frequencies and reducing time series lengths on the accuracy of estimating dynamical characteristics from gPPG and rPPG signals. Results demonstrate that a sampling frequency of 200 Hz provides an optimal balance, maintaining robust correlations in dynamical indices while reducing computational load. Furthermore, analysis of varying time series lengths revealed that the dynamical properties stabilize sufficiently at around 170 s, achieving an error of less than 5%. A comparative analysis between gPPG and rPPG revealed no significant statistical differences, confirming their similar effectiveness in estimating dynamical properties under controlled conditions. These results enhance the reliability and applicability of PPG-based health monitoring technologies. Full article

This paper presents a real-time smart mirror system combining multiple AI modules for multimodal health monitoring. The proposed platform integrates three core components: facial expression analysis, remote photoplethysmography (rPPG), and conversational AI. A key innovation lies in transforming the Moving Average rPPG (MA-rPPG) model—originally developed for offline batch processing—into a real-time, continuously streaming setup, enabling seamless heart rate and peripheral oxygen saturation (SpO₂) monitoring using standard webcams. The system also incorporates the DeepFace facial analysis library for live emotion, age detection, and a Generative Pre-trained Transformer 4o (GPT-4o)-based mental health chatbot with bilingual (English/Korean) support and voice synthesis. Embedded into a touchscreen mirror with Graphical User Interface (GUI), this solution delivers ambient, low-interruption interaction and real-time user feedback. By unifying these AI modules within an interactive smart mirror, our findings demonstrate the feasibility of integrating multimodal sensing (rPPG, affect detection) and conversational AI into a real-time smart mirror platform. This system is presented as a feasibility-stage prototype to promote real-time health awareness and empathetic feedback. The physiological validation was limited to a single subject, and the user evaluation constituted only a small formative assessment; therefore, results should be interpreted strictly as preliminary feasibility evidence. The system is not intended to provide clinical diagnosis or generalizable accuracy at this stage. Full article

This study investigates the effects of heart rate and arm tremor on performance in elite archery, using non-contact physiological monitoring from real Olympic competition footage. A total of 50 video segments were extracted from publicly available international broadcasts, comprising athletes of various backgrounds. From these, heart rate signals were estimated via remote photoplethysmography (rPPG) from facial regions, and micro-movements were quantified from right and left arm regions using feature point tracking. Ordinal logistic regression was employed to evaluate the relationship between biometric variables and archery scores (10, 9, ≤8 points). Results showed that elevated heart rate ($\beta$ = −0.1166; $p<$ 0.001) and greater right-arm movement ($\beta$ = −6.1747; p = 0.008) were significantly associated with lower scores. Athletes scoring 10 points exhibited significantly lower heart rate ($p<$ 0.001) and reduced right-arm tremor (p = 0.010) compared to others. These findings support the hypothesis that physiological arousal and biomechanical instability impair performance, and they further demonstrate the feasibility of contactless monitoring in real competition environments. The proposed method enables objective, in-game performance evaluation and supports the development of personalized training systems for precision sports. Full article

Objective: In this paper, we propose a novel approach to heart rate (HR) detection based on the evaluation of oscillatory patterns of continuous wavelet transform as a method of time-frequency analysis. HR detection based on electrocardiogram (ECG) or photoplethysmogram (PPG) signals can be performed using the same technique. Methods: The developed approach was tested on ECG (lead V1) and PPG (standard recording on the ring finger of the left hand and differential signal) for 10 min in 40 generally healthy volunteers (aged 26.8 ± 3.22 years). A comparison was made with the traditional HR detection method based on R-peak shape analysis. Results: Based on a number of statistical evaluations, the comparison yielded an acceptable degree of agreement between the results of the proposed method and the traditional method (the discrepancy between the results did not exceed 3.41%). The distortion of the signal shape and its noise do not affect the quality of HR detection by the proposed method; so, additional filtering or changes in the implemented algorithm are not required, as demonstrated by processing both the differential PPG signal and the PPG signals recorded during the patient’s walking. Conclusions: The proposed method allows obtaining HR information with a higher equidistant sampling frequency and expansion of the information on the frequency composition of HRV. Full article