Search Results (41)

This paper proposes a two-stage hybrid framework for biosignal quality validation that produces beat-level or segment-level labels for real-time filtering and offline dataset curation. The framework is quantitatively validated exclusively on ECG data. Its modular architecture is designed to extend to further non-stationary periodic biomedical time-series signals including photoplethysmography (PPG), impedance cardiography (ICG), phonocardiography (PCG), electromyography (EMG), and electroencephalography (EEG) through modality-specific parameter adaptation; however, this broader applicability currently reflects architectural extensibility rather than experimentally validated performance. A prerequisite is synchronized acquisition of the primary biosignal together with inertial motion sensing (IMU/accelerometer) and electrode impedance or lead-off status, with the IMU positioned near the sensing electrodes. The first stage performs sensor-integrity gating to reject intervals corrupted by motion or poor electrode contact. The second stage applies software signal quality indices to the remaining beats, including physiological plausibility constraints (R to R peaks analysis), DTW-based morphological consistency against adaptive templates, frequency domain SNR estimation, and baseline wander quantification. This study systematically evaluates and compares the classification performance of six complementary sensor-level and software-based signal quality assessment methods. When integrated within the proposed hybrid framework, validation against expert-annotated ECG quality labels from 20 healthy participants demonstrates high methodological classification accuracy (98.1%), achieving approximately a 98% F1-score, 99% sensitivity, and 97% specificity. Prospective validation on patient populations with cardiovascular pathology is identified as a necessary step toward clinical deployment. This modular approach improves the reliability of downstream analysis by preventing corrupted data from entering feature extraction and model training pipelines, enabling more stable physiological monitoring in free-living conditions, reducing false alarms in continuous monitoring applications, and generating higher-quality datasets for AI-based diagnostic systems. Full article

Thedeep score-based diffusion (DeScoD) model performs well in electrocardiogram (ECG) denoising tasks. However, due to the theoretical error lower bound in approximating functions with linear transformations, it often lacks flexibility when fitting non-stationary noise, baseline wander, or morphologically variable features such as QRS complexes in ECG signals. In this paper, we propose a Kolmogorov–Arnold network enhanced deep score-based diffusion (KAN-DeScoD) model, which is the first to integrate Kolmogorov–Arnold network (KAN) layers into an ECG denoising diffusion model. By leveraging KAN’s adaptive activation functions, which more finely capture the complex structures within ECG signals, the model’s robustness in high-noise environments, as well as the accuracy and stability of signal reconstruction, are improved. We validate the effectiveness of the proposed method on the QT Database and the MIT-BIH Noise Stress Test Database (NSTDB). Experimental results show that under different shots and noise intensities, ours outperforms the DeScoD model across multiple metrics. The research results demonstrate the effectiveness of introducing KAN, which improves the model’s robustness in high-noise environments and the accuracy of signal reconstruction. Full article

Background: Attention-Deficit/Hyperactivity Disorder (ADHD) has been less frequently and extensively investigated in university students than in children, despite substantial evidence demonstrating its significant impact on academic performance and negative affect, such as anxiety. We conducted two studies to address this gap. Methods: The objective of our first study (n = 233) was to assess the prevalence of subclinical ADHD among medical students and examine its associations with comorbid mental health conditions, such as Depression, Anxiety and Stress (DASS-21). In the second pilot intervention study (n = 16), we compared the ratio of negative and positive emotions (PANAS) and anxiety (STAI-S-5) before and after neurofeedback-based relaxation training in two groups of students: one with high scores and another with low scores on the Adult ADHD Self-Report Scale (ASRS). Results: According to our results, more than 50% of students showed risk for ADHD symptoms, and linear regression analyses revealed a strong association between ADHD symptoms and the prevalence of negative affect. Interestingly, no significant differences were found in ADHD and DASS scale scores between students who were falling behind and those progressing in line with the curriculum. Further results of the second study were inconclusive in several areas. In the examined group, a significant increase was observed in one of the core symptoms of ADHD—mind wandering—by the end of the intervention, compared to the baseline. Additionally, frustration levels were significantly higher at the second measurement point among participants with higher ASRS scores. Conclusions: Compared to the literature, it can be concluded that while longer interventions tend to be effective, two sessions are insufficient to reduce symptom. Full article

This paper proposes an adaptive layer-dependent threshold function (ALDTF) for denoising electrocardiogram (ECG) and multimode optical fiber-based cardiopulmonary signals. Based on wavelet transform, the method employs a layer-dependent threshold function strategy that utilizes the non-zero periodic peak (NZOPP) of the signal’s normalized autocorrelation function to adaptively determine the optimal threshold for each decomposition layer. The core idea applies soft thresholding at lower layers (high-frequency noise) to suppress pseudo-Gibbs oscillations, and hard thresholding at higher layers (low-frequency noise) to preserve signal amplitude and morphology. The experimental results show that for ECG signals contaminated with baseline wander (BW), electrode motion (EM) artifacts, muscle artifacts (MA), and mixed (MIX) noise, ALDTF outperforms existing methods—including SWT, DTCWT, and hybrid approaches—across multiple metrics. It achieves a ΔSNR improvement of 1.68–10.00 dB, ΔSINAD improvement of 1.68–9.98 dB, RMSE reduction of 0.02–0.56, and PRD reduction of 2.88–183.29%. The method also demonstrates excellent performance on real ECG and optical fiber cardiopulmonary signals, preserving key diagnostic features like QRS complexes and ST segments while effectively suppressing artifacts. ALDTF provides an efficient, versatile solution for physiological signal denoising with strong potential in wearable real-time monitoring systems. Full article

The ST segment of an ECG signal is a feature that changes in the event of cardiac ischemia, a condition that is an early warning sign of myocardial infarction. Being able to monitor this feature in real time would be highly beneficial for preventing recurrent heart attacks. However, to be worn daily, such a monitoring device must be extremely miniaturized, down to the scale of a single integrated circuit. Currently, it is possible to integrate a heart rate detector, but, to our knowledge, no existing work presents a chip capable of detecting ST segment deviation. This is mainly because accurate ST segment measurement requires low-distortion signal processing, as specified in the International Electrotechnical Commission (IEC) standard. At the same time, the system is required to filter out baseline wander, whose frequency components may partially overlap with those of the ST segment. In this study, we relied on wavelet-based analysis and reconstruction to compare several wavelet types. We optimized their hyperparameters to minimize implementation complexity while satisfying the low-distortion constraints. We also propose an ASIC-oriented architecture and evaluate its post-layout performance in terms of area and power consumption. The post-layout results indicate that the Daubechies wavelet db3 offers the best trade-off among the evaluated configurations. It exhibits an area utilization of $1.18$ mm² and a post-layout power consumption of $4.89$ $\mathsf{\mu }$W, while preserving the ST segment in compliance with the IEC standard, thanks in particular to its effective baseline wandering filtering of $6.9$ dB. These results demonstrate the feasibility of embedding automatic ST segment extraction on-chip. Full article

Pedestrian panic behavior is a primary cause of overcrowding and stampede accidents in public micro-road network areas with high pedestrian density. However, reliably detecting such behaviors remains challenging due to their inherent complexity, variability, and stochastic nature. Current detection models often rely on single-modality features, which limits their effectiveness in complex and dynamic crowd scenarios. To overcome these limitations, this study proposes a contour-driven multimodal framework that first employs a CNN (CDNet) to estimate density maps and, by analyzing steep contour gradients, automatically delineates a candidate panic zone. Within these potential panic zones, pedestrian trajectories are analyzed through LSTM networks to capture irregular movements, such as counterflow and nonlinear wandering behaviors. Concurrently, semantic recognition based on Transformer models is utilized to identify verbal distress cues extracted through Baidu AI’s real-time speech-to-text conversion. The three embeddings are fused through a lightweight attention-enhanced MLP, enabling end-to-end inference at 40 FPS on a single GPU. To evaluate branch robustness under streaming conditions, the UCF Crowd dataset (150 videos without panic labels) is processed frame-by-frame at 25 FPS solely for density assessment, whereas full panic detection is validated on 30 real Itaewon-Stampede videos and 160 SUMO/Unity simulated emergencies that include explicit panic annotations. The proposed system achieves 91.7% accuracy and 88.2% F1 on the Itaewon set, outperforming all single- or dual-modality baselines and offering a deployable solution for proactive crowd safety monitoring in transport hubs, festivals, and other high-risk venues. Full article

Achieving both accurate and real-time monitoring heartbeat signals by non-invasive sensing techniques is challenging due to various noise interferences. In this paper, we propose an enhanced discrete wavelet transform (DWT) method that incorporates objective denoising quality assessment metrics to determine accurate thresholds and adaptive threshold functions. Our approach begins by denoising ECG signals from various databases, introducing several types of typical noise, including additive white Gaussian (AWG) noise, baseline wandering noise, electrode motion noise, and muscle artifacts. The results show that for Gaussian white noise denoising, the enhanced DWT can achieve 1–5 dB SNR improvement compared to the traditional DWT method, while for real noise denoising, our proposed method improves the SNR tens or even hundreds of times that of the state-of-the-art denoising techniques. Furthermore, we validate the effectiveness of the enhanced DWT method by visualizing and comparing the denoising results of heartbeat signals monitored by fiber-optic micro-vibration sensors against those obtained using other denoising methods. The improved DWT enhances the quality of heartbeat signals from non-invasive sensors, thereby increasing the accuracy of cardiovascular disease diagnosis. Full article

Seismocardiography (SCG) and Gyrocardiography (GCG) use lightweight, miniaturized accelerometers and gyroscopes to record, respectively, cardiac-induced linear accelerations and angular velocities of the chest wall. These inertial sensors are also sensitive to thoracic movements with respiration, which cause baseline wanderings in SCG and GCG signals. Nowadays, accelerometers and gyroscopes are widely integrated into smartphones, thus increasing the potential of SCG and GCG as cardiorespiratory monitoring tools. This study investigates the accuracy of smartphone inertial sensors in simultaneously measuring instantaneous heart rates and breathing rates. Smartphone-derived SCG and GCG signals were acquired from 10 healthy subjects at rest. The performances of heartbeats and respiratory acts detection, as well as of inter-beat intervals (IBIs) and inter-breath intervals (IBrIs) estimation, were evaluated for both SCG and GCG via the comparison with simultaneous electrocardiography and respiration belt signals. Heartbeats were detected with a sensitivity and positive predictive value (PPV) of 89.3% and 93.3% in SCG signals and of 97.3% and 97.9% in GCG signals. Moreover, IBIs measurements reported strong linear relationships (R² > 0.999), non-significant biases, and Bland–Altman limits of agreement (LoA) of ±7.33 ms for SCG and ±5.22 ms for GCG. On the other hand, respiratory acts detection scored a sensitivity and PPV of 95.6% and 94.7% for SCG and of 95.7% and 92.0% for GCG. Furthermore, high R² values (0.976 and 0.968, respectively), non-significant biases, and an LoA of ±0.558 s for SCG and ±0.749 s for GCG were achieved for IBrIs estimates. The results of this study confirm that smartphone inertial sensors can provide accurate measurements of both instantaneous heart rate and breathing rate without the need for additional devices. Full article

This review systematically analyzes the latest advancements in preprocessing techniques for Electrocardiography (ECG) and Magnetocardiography (MCG) signals over the past decade. ECG and MCG play crucial roles in cardiovascular disease (CVD) detection, but both are susceptible to noise interference. This paper categorizes and compares different ECG denoising methods based on noise types, such as baseline wander (BW), electromyographic noise (EMG), power line interference (PLI), and composite noise. It also examines the complexity of MCG signal denoising, highlighting the challenges posed by environmental and instrumental interference. This review is the first to systematically compare the characteristics of ECG and MCG signals, emphasizing their complementary nature. MCG holds significant potential for improving the precision of CVD clinical diagnosis. Additionally, it evaluates the limitations of current denoising methods in clinical applications and outlines future directions, including the potential of explainable neural networks, multi-task neural networks, and the combination of deep learning with traditional methods to enhance denoising performance and diagnostic accuracy. In summary, while traditional filtering techniques remain relevant, hybrid strategies combining machine learning offer substantial potential for advancing signal processing and clinical diagnostics. This review contributes to the field by providing a comprehensive framework for selecting and improving denoising techniques, better facilitating signal quality enhancement and the accuracy of CVD diagnostics. Full article

Electrocardiographic signals (ECG) are ubiquitous, which justifies the research of their optimal storage and transmission. However, proposals for non-uniform signal sampling must take into account the priority of diagnostic data accuracy and record integrity, as well as robustness to noise and interference. In this study, two novel methods are introduced, each utilizing a distinct neural network architecture for optimizing non-uniform sampling of ECG signal. A transformer model refines each time point selection through an iterative process using gradient descent optimization, with the goal of minimizing the mean squared error between the original and resampled signals. It adaptively modifies time points, which improves the alignment between both signals. In contrast, the Temporal Convolutional Network model trains on the original signal, and gradient descent optimization is utilized to improve the selection of time points. Evaluation of both strategies’ efficacy is performed by calculating signal distances at lower and higher sampling rates. First, a collection of synthetic data points that resembled the P-QRS-T wave was used to train the model. Then, the ECG-ID database for real data analysis was used. Filtering to remove baseline wander followed by evaluation and testing were carried out in the real patient data. The results, in particular MSE = 0.0005, RMSE = 0.0216, and Pearson’s CC = 0.9904 for 120 sps in the case of the transformer patient data model, provide viable paths for maintaining the precision and dependability of ECG-based diagnostic systems at much lower sampling rate. Outcomes indicate that both techniques are effective at improving the fidelity between the original and modified ECG signals. Full article

This paper presents a novel approach to mind-wandering prediction in the context of webcam-based online learning. We implemented a Singular Value Decomposition (SVD)-based 1D temporal eye-signal extraction method, which relies solely on eye landmark detection and eliminates the need for gaze tracking or specialized hardware, then extract suitable features from the signals to train the prediction model. Our thorough experimental framework facilitates the evaluation of our approach alongside baseline models, particularly in the analysis of temporal eye signals and the prediction of attentional states. Notably, our SVD-based signal captures both subtle and major eye movements, including changes in the eye boundary and pupil, surpassing the limited capabilities of eye aspect ratio (EAR)-based signals. Our proposed model exhibits a 2% improvement in the overall Area Under the Receiver Operating Characteristics curve (AUROC) metric and 7% in the F1-score metric for ‘not-focus’ prediction, compared to the combination of EAR-based and computationally intensive gaze-based models used in the baseline study These contributions have potential implications for enhancing the field of attentional state prediction in online learning, offering a practical and effective solution to benefit educational experiences. Full article

The trend of using deep learning techniques to classify arbitrary tasks has grown significantly in the last decade. Such techniques in the background provide a stack of non-linear functions to solve tasks that cannot be solved in a linear manner. Naturally, deep learning models can always solve almost any problem with the right amount of functional parameters. However, with the right set of preprocessing techniques, these models might become much more accessible by negating the need for a large set of model parameters and the concomitant computational costs that accompany the need for many parameters. This paper studies the effects of such preprocessing techniques, and is focused, more specifically, on the resulting learning representations, so as to classify the arrhythmia task provided by the ECG MIT-BIH signal dataset. The types of noise we filter out from such signals are the Baseline Wander (BW) and the Powerline Interference (PLI). The learning representations we use as input to a Convolutional Neural Network (CNN) model are the spectrograms extracted by the Short-time Fourier Transform (STFT) and the scalograms extracted by the Continuous Wavelet Transform (CWT). These features are extracted using different parameter values, such as the window size of the Fourier Transform and the number of scales from the mother wavelet. We highlight that the noise with the most significant influence on a CNN’s classification performance is the BW noise. The most accurate classification performance was achieved using the 64 wavelet scales scalogram with the Mexican Hat and with only the BW noise suppressed. The deployed CNN has less than 90k parameters and achieved an average F1-Score of 90.11%. Full article

The electrocardiogram (ECG) is widely used in medicine because it can provide basic information about different types of heart disease. However, ECG data are usually disturbed by various types of noise, which can lead to errors in diagnosis by doctors. To address this problem, this study proposes a method for denoising ECG based on disentangled autoencoders. A disentangled autoencoder is an improved autoencoder suitable for denoising ECG data. In our proposed method, we use a disentangled autoencoder model based on a fully convolutional neural network to effectively separate the clean ECG data from the noise. Unlike conventional autoencoders, we disentangle the features of the coding hidden layer to separate the signal-coding features from the noise-coding features. We performed simulation experiments on the MIT-BIH Arrhythmia Database and found that the algorithm had better noise reduction results when dealing with four different types of noise. In particular, using our method, the average improved signal-to-noise ratios for the three noises in the MIT-BIH Noise Stress Test Database were 27.45 db for baseline wander, 25.72 db for muscle artefacts, and 29.91 db for electrode motion artefacts. Compared to a denoising autoencoder based on a fully convolutional neural network (FCN), the signal-to-noise ratio was improved by an average of 12.57%. We can conclude that the model has scientific validity. At the same time, our noise reduction method can effectively remove noise while preserving the important information conveyed by the original signal. Full article

This research article explores a hybrid strategy that combines an adaptive iterative filtering (IF) method and the fast discrete lifting-based wavelet transform (LWT) to eliminate power-line noise (PLI) and baseline wander from an electrocardiogram (ECG) signal. Due to its correct mathematical basis and its guaranteed a priori convergence, the iterative filtering approach was preferred over empirical mode decomposition (EMD). The noisy modes generated from the IF are fed to an LWT system so as to be disintegrated into the detail and the approximation coefficients. These coefficients are then scaled using a threshold method to generate a noise-free signal. The proposed strategy improves the quality and allows us to precisely preserve the vital components of the signal. The method’s potency has been established empirically by calculating the improvement in signal-to-noise ratio, cross-correlation coefficient and percent root-mean-square difference for different recordings available on the MIT-BIH arrhythmia database and then compared to numerous existing methods. Full article

The high mortality rate associated with cardiac abnormalities highlights the need of accurately detecting heart disorders in the early stage so to avoid severe health consequence for patients. Health trackers have become popular in the form of wearable devices. They are aimed to perform cardiac monitoring outside of medical clinics during peoples’ daily lives. Our paper proposes a new diagnostic algorithm and its implementation adopting a FPGA-based design. The conceived system automatically detects the most common arrhythmias and is also able to evaluate QT-segment lengthening and pulmonary embolism risk often caused by myocarditis. Debug and simulations have been carried out firstly in Matlab environment and then in Quartus IDE by Intel. The hardware implementation of the embedded system and the test for the functional accuracy verification have been performed adopting the DE1_SoC development board by Terasic, which is equipped with the Cyclone V 5CSEMA5F31C6 FPGA by Intel. Properly modified real ECG signals corrupted by a mixture of muscle noise, electrode movement artifacts, and baseline wander are used as a test bench. A value of 99.20% accuracy is achieved by taking into account 0.02 mV for the root mean square value of noise voltage. The implemented low-power circuit is suitable as a wearable decision support device. Full article