ECG Signal Research Articles

Retraction notice: Machine learning based pervasive analytics for ECG signal analysis

Retraction notice: Machine learning based pervasive analytics for ECG signal analysis

An ECG Signal Classification System Using EMD and Pipelined Yolo Network

An ECG Signal Classification System Using EMD and Pipelined Yolo Network

Automatic Diagnosis of Sleep Apnea and Its Type Using Linear Discriminant Analysis and Support Vector Machine Based on the Characteristics of Two ECG and PPG Signals

Automatic Diagnosis of Sleep Apnea and Its Type Using Linear Discriminant Analysis and Support Vector Machine Based on the Characteristics of Two ECG and PPG Signals

Improving automated labeling with deep learning and signal segmentation for accurate ECG signal analysis

Improving automated labeling with deep learning and signal segmentation for accurate ECG signal analysis

Potential applications of humanoid robotic touch for social regulation of emotion: evidence from ECG and fNIRS

ABSTRACT Human-robot touch interaction plays an essential role in emotional support and human mental health support. The role of therapeutic robots such as Paro in mental support has been widely investigated. However, the impact of humanoid robotic touch on the social regulation of users’ emotions is still unknown. Therefore, a mixed experiment was conducted with the type of touch (grip versus contact) as the between-subjects factor and the presence of touch during movie reception (with versus without touch) as the within-subjects factor. The subjective perception of emotion, ECG, and fNIRS signals were collected during the experiment. The results showed that robot touch regulates subjectively positive emotions, reduces HR, increases HRV, and helps suppress the brain activity on the right DLPFC. No main effect of touch type was found on the regulation effect of subjective emotions, autonomic responses, and central nervous responses. The study provides subjective and neurophysiological evidence for the great potential of humanoid robotics for the social regulation of emotion.

Advances in ECG and PCG-based cardiovascular disease classification: a review of deep learning and machine learning methods

Cardiovascular diseases (CVD) have been found to be prevalent in society, frequently ending in death. According to the findings of a recent survey, the mortality rate is increasing due to the prevalence of adult cigarette consumption, elevated blood pressure, high cholesterol levels, and obesity. The previously mentioned causes are exacerbating the severity of the condition. A pressing necessity exists for a study on the variability of these factors and their impact on cardiovascular disease (CVD). This involves the use of advanced tools to detect the disease early on and aid in the reduction of fatality rates. With their extensive methodologies that would help in the early CVD prediction and recognition of behavioral patterns in large amounts of data, artificial intelligence, and data mining disciplines offer a broad study potential. The results of these predictions will help physicians make decisions and early diagnoses, decreasing the risk of patient death. This work compares and reports the classification, machine learning, and deep learning algorithms that predict cardiovascular illnesses. For this study, articles from 2012 to 2023 were considered; after filtering, 82 articles were chosen for primary research. Future researchers will benefit from this review on cardiovascular disorders by better understanding the Deep Learning and Machine Learning models now in the healthcare sector. The review encompasses commonly employed methodologies such as support vector machine, decision tree, random forest, and convolutional neural networks (CNNs). Additionally, this survey aggregates and presents information on the performance metrics used to report accuracy. It also goes over the most popular datasets used by various diagnostic models (ECG and PCG signals datasets). In addition, it emphasizes prominent publishers, journals, and conferences that serve as platforms for the evaluation of scholarly works. Additionally, it will facilitate their understanding of the unresolved challenges or hurdles experienced by past researchers. A lack of more extensive and consistent datasets was the most common issue, followed by the need to improve existing models.

Deep learning hybrid model ECG classification using AlexNet and parallel dual branch fusion network model.

Cardiovascular diseases are a cause of death making it crucial to accurately diagnose them. Electrocardiography plays a role in detecting heart issues such as heart attacks, bundle branch blocks and irregular heart rhythms. Manual analysis of ECGs is prone to mistakes and time consuming, underscoring the importance of automated methods. This study uses AI models like AlexNet and a dual branch model for categorizing ECG signals from the PTB Diagnostic ECG Database. AlexNet achieved a validation accuracy of 98.64% and a test set accuracy of 99% while the dual branch fusion network model achieved a test set accuracy of 99%. Data preprocessing involved standardizing, balancing and reshaping ECG signals. These models exhibited precision, sensitivity and specificity. In comparison to state of the arts' models such as Hybrid AlexNet SVM and DCNN LSTM our proposed models displayed performance. The high accuracy rates of 99% underscore their potential for ECG classification. These results validate the advantages of incorporating learning models into setups for automated ECG analysis providing adaptable solutions for various healthcare settings including rural areas.

Continuous QTc monitoring for patients intoxicated with QTc prolonging medication in the ED: a proof of principle.

Background: To determine when patients presenting with an overdose of QT-interval prolonging drugs can be discharged safely can be challenging, especially when the moment of intoxication and/or the substance(s) ingested are unknown. Objective: In a proof of principle study, we aimed to study if continuous QTc interval analysis can be used to establish optimal observation duration of patients with an intoxication with QTc prolonging medication. Methods: For patients presenting with an intoxication with QT-interval prolonging drugs in the emergency department (ED), ECG signals sampled at 500Hz were pre-processed, and the average heart rate corrected QT-interval (QTc) per 5 minutes was calculated and plotted against time. A third order polynomial was fitted to visualize when the QTc would be highest (the electrophysiological Tmax). This point in time was compared to the estimated Tmax based on pharmacokinetic properties of the ingested substance. Results: In a retrospective biobank-based study, a total of 22 ED visits (of 15 patients) were analyzed. An electrophysiological Tmax could be calculated for 17/22 visits. The remaining five patients presented either long after the electrophysiological Tmax (n=4) or were admitted to the ward before reaching the Tmax (n=1). The average (SD) difference between the estimated Tmax based on drug properties and the calculated electrophysiological Tmax was 18 (133) minutes (range -158-296 minutes). Despite the wide range, there was a significant correlation between recorded electrophysiological Tmax and estimated Tmax (r=0.67, p=0.012). Conclusion: Continuous electrophysiological monitoring can be used as an adjunct to determine the toxicokinetic Tmax for patients presenting with an intoxication, especially when the time of ingestion or the substance ingested are unknown.

Knowledge-enhanced meta-transfer learning for few-shot ECG signal classification

Knowledge-enhanced meta-transfer learning for few-shot ECG signal classification

Adaptive ECG Signal Denoising Algorithm Based on the Improved Whale Optimization Algorithm

Adaptive ECG Signal Denoising Algorithm Based on the Improved Whale Optimization Algorithm

Advanced Noise-Resistant Electrocardiography Classification Using Hybrid Wavelet-Median Denoising and a Convolutional Neural Network.

The classification of ECG signals is a critical process because it guides the diagnosis of the proper treatment process for the patient. However, any form of disturbance with ECG signals can be highly conspicuous because of the mechanics involved in data acquisition from living beings, which has a significant impact on the classification procedure. The purpose of this research work is to advance ECG signal classification results by employing numerous denoising methods and, in turn, boost the accuracy of cardiovascular diagnoses. To simulate realistic conditions, we added various types of noise to ECG data, including Gaussian, salt and pepper, speckle, uniform, and exponential noise. To overcome the interference of noise from environments in the obtained ECG signals, we employed wavelet transform, median filter, Gaussian filter, and the hybrid of the wavelet and median filters. The proposed hybrid denoising method has better results than the other methods because of the use of wavelet multi-scale analysis and the ability of the median filter to avoid the loss of vital ECG characteristics. Thus, despite a certain proximity in the values, the hybrid method is significantly more accurate and reliable, as evidenced by the mean squared error (MSE), mean absolute error (MAE), R-squared, and Pearson correlation coefficient. More specifically, the hybrid approach provided an MSE of 0.0012 and an MAE of 0.025, the R-squared value for this study was 0.98, and the Pearson correlation coefficient was 0.99, which provides a very good resemblance to the original ECG confirmation. The proposed classification model is based on the modified lightweight CNN or MLCNN that was trained using the noisy and the denoised data. The findings demonstrated that by applying the denoised data, the testing accuracy, precision, recall, and F1 scores achieved 0.92, 0.91, 0.90, and 0.91 for the datasets, while the noisy data achieved 0.80, 0.78, 0.82, and 0.80, respectively. In this study, the signal quality and denoising methods were found to enhance ECG signal classification and diagnostic accuracy while encouraging proper preprocessing in future studies and applications for real-time ECG for cardiac care.

Enhanced CAD Detection Using Novel Multi-Modal Learning: Integration of ECG, PCG, and Coupling Signals

Early and highly precise detection is essential for delaying the progression of coronary artery disease (CAD). Previous methods primarily based on single-modal data inherently lack sufficient information that compromises detection precision. This paper proposes a novel multi-modal learning method aimed to enhance CAD detection by integrating ECG, PCG, and coupling signals. A novel coupling signal is initially generated by operating the deconvolution of ECG and PCG. Then, various entropy features are extracted from ECG, PCG, and its coupling signals, as well as recurrence deep features also encoded by integrating recurrence plots and a parallel-input 2-D CNN. After feature reduction and selection, final classification is performed by combining optimal multi-modal features and support vector machine. This method was validated on simultaneously recorded standard lead-II ECG and PCG signals from 199 subjects. The experimental results demonstrate that the proposed multi-modal method by integrating all signals achieved a notable enhancement in detection performance with best accuracy of 95.96%, notably outperforming results of single-modal and joint analysis with accuracies of 80.41%, 86.51%, 91.44%, and 90.42% using ECG, PCG, coupling signal, and joint ECG and PCG, respectively. This indicates that our multi-modal method provides more sufficient information for CAD detection, with the coupling information playing an important role in classification.

Detection of ATTR cardiac amyloidosis using a novel artificial intelligence algorithm for wearable-adapted noisy single-lead electrocardiograms

Abstract Background Amyloid cardiomyopathy (ATTR-CM) is often underdiagnosed, with very few diagnosed before the onset of symptoms, leading to delayed use of disease-modifying therapies that could have substantial prognostic implications early in the disease course. In the US, the disease has a substantially higher prevalence among Black adults, who are further challenged by lower access to healthcare services and lower preventative care, further reducing the ability for clinician-led diagnosis. We sought to develop an approach for identifying ATTR-CM on noisy single-lead ECGs obtainable on portable and wearable ECG devices, allowing broad use in community screening. Purpose To develop a portable/wearable device-adapted AI-ECG algorithm capable of detecting ATTR-CM on noisy single-lead ECGs. Methods We identified patients with ATTR-CM from 2015 through June 2023 across 5 hospitals of a large U.S.-based hospital system using positive bone scintigraphy scans or pharmacotherapy with an approved transthyretin stabilizer. In the development cohort, we used lead I of 12 lead ECGs, commonly captured by portable and wearable ECG devices, to train a novel noise-adapted model explicitly designed to infer information against simulated real-world noises. For this, ECG signals were augmented using random Gaussian real-world noise within four frequency ranges at multiple signal-to-noise ratios. We tested the model in an independent set of cases and matched controls drawn from July through December 2023. Results The development cohort consisted of 1,011 ECGs from 234 patients (median age 79 years [IQR:70-85], 17% women), with 10:1 age- and sex-matched 10,110 controls with 10,110 ECGs (age 79 [IQR:70-85] years 17.7% female). In the independent test set of 139 ECGs from 47 patients (age 80 [75-86] years, 32% women) and matched controls (1390 ECGs and patients) from July through December 2023, the AUROC (area under the receiver operating characteristic curve) of the AI-ECG model for ATTR-CM was 0.90 [95%CI: 0.88-0.92] (A). The AUPRC was 0.44 [0.36-0.53]. The sensitivity and specificity of the model were 0.85 and 0.80. The positive predictive value ranged from 0.12 at a 3% prevalence level to 0.59 at a 25% prevalence (B). Conclusions A novel portable and wearable-devices adapted model demonstrates promising performance for detecting ATTR-CM on single-lead ECGs, representing a more accessible method for screening in community-dwelling adults.

Digitisation of electrocardiographic images enable artificial intelligence-based mortality prediction

Abstract Background Artificial intelligence-based electrocardiogram (AI-ECG) models have shown excellent performance in a vast number of diagnostic and prognostic tasks. One of such tasks, which is critical for patient risk stratification and clinical intervention, is the prediction of adverse events, including death. Although a plethora of existing AI-ECG models have shown their accuracy in mortality prediction, their development and utilization currently rely on digital ECG signals for training and application. Purpose Many hospitals around the world have a significant number of ECGs stored in image formats or printed as paper ECGs. To address this problem, we developed and tested an ECG image digitisation pipeline to extract digital traces from ECG images and compare the performance of AI-ECG models for mortality prediction, under both natively digital and digitised signal formats. Methods A secondary care dataset of over 1 million ECGs (1,163,401 ECGs; 189,539 patients) was used for this analysis, in which both a natively digital signal and a PDF image was available. The ECG digitisation pipeline comprised of coloured PDF image inputs (2200 x 1700), initially processed to extract the ECG traces and grid features. ECG signals were extracted from the traces after cropping/restoration of the images and scaled appropriately based on the detected grid (2.5-sec strips per lead, 0.5 mV per grid block). Subsequently, both digital and digitised signals were identically pre-processed in order to acquire the ECG data fed into the AI model. In particular, a previously successful convolutional neural network architecture with a discrete-time survival loss function was used for mortality prediction. The model was trained under multiple configurations, using either the digital or the digitised ECGs as training/testing datasets (Figure 1). ECGs were split by patient ID to ensure cross-subject generalization with a training/validation/testing split of 50%/10%/40% ratio, respectively. Finally, the models were evaluated by the c-index. Results A comparison of the digital and digitised ECGs on the respective 2.5-sec strips available in both formats revealed a mean absolute error of 0.014 (CI: 0.07 – 0.021) and a Pearson’s R of 0.98 (CI: 0.96 – 1.00) between the signals, demonstrating an accurate ECG digitisation (Figure 2 shows overlapping natively digital (black) and digitised (red) traces). The performance of the digitisation was also evident in the performance of the AI-ECG mortality prediction task, with the model achieving c-indices of 0.775 (CI: 0.774 – 0.776) and 0.772 (CI: 0.771 – 0.774) when trained and tested on digital and digitised signals, respectively. Conclusion We described the first digitisation-based AI-ECG model for mortality prediction. Our results demonstrate the potential for clinical settings that lack the digital ECG infrastructure, to develop and deploy AI models using ECG image formats.Figure 1Figure 2

Cardiac autonomic dysfunction in Takotsubo cardiomyopathy: an observational study

Abstract Takotsubo syndrome (TTS), commonly perceived as a benign and reversible condition has received attention due to emerging registry data revealing its prognosis to be comparable to acute coronary syndrome (ACS). Following ACS, a notable subset of patients develops autonomic dysfunction, whith unfavourable prognostic implications. Periodic repolarization dynamics (PRD) and deceleration capacity (DC), derived from ECG signals, are parameters capable of quantifying cardiac autonomic function. In this study, we aimed to assess autonomic dysfunction in patients with TTC in comparison to ST-elevation myocardial infarction (STEMI) patients. Consecutive patients diagnosed with TTS were prospectively recruited into an observational, single-centre cohort study. All patients underwent a 30-minute high-resolution ecg recording. Subsequent recordings were conducted at 4 and 12 months post-acute event. PRD and DC as markers for sympathetic and parasympathetic activity, respectively, were assessed using established methods. The control group comprised patients with STEMI. Statistical comparisons between groups were performed using the Mann-Whitney U-test, with a corrected significance level for multiple testing of α< .017. Between July 2021 and December 2023, 57 patients diagnosed with TTS were recruited (interquartile range [IQR]) 69.0 years (62.0-78.0), 98.3% women. A control group was derived from a pre-existing cohort of STEMI patients, matched through propensity score matching to adjust for age and sex (median age [IQR]: 69.0 [59.0-76.0] years, 98.3% women). At baseline, there was no significant difference in PRD between TTS patients and controls (median [IQR]: TTS: 5.28deg² [3.43-9.93], STEMI: 4.40deg² [2.24-7.21]; p = .04). However, at 4 months post-acute event, PRD was notably higher in TTS patients compared to controls (median [IQR]: TTS: 5.20deg² [2.71-8.17], STEMI: 2.63deg² [1.86-4.92]; p = .011). Interestingly, at 12-month follow-up, PRD in TTS patients did not differ from that of controls (median [IQR]: TTS: 4.55deg² [2.89-8.40], STEMI: 4.15deg² [2.42-6.85]; p = .68). Regarding DC, no significant differences were observed between TTS and STEMI patients at any time point (baseline: median [IQR]: TTS: 3.95 ms [2.27-5.51], STEMI: 3.52 ms [2.30-7.24], p = .51; 4 months: median [IQR]: TTS: 6.63 ms [4.59-8.64], STEMI: 7.41 ms [4.63-9.25], p = .39; 12 months: median [IQR]: TTS: 4.98 ms [4.19-6.67], STEMI: 7.48 ms [4.92-10.28], p = .17). Patients with TTS show substantial signs of cardiac autonomic dysfunction, similar or even higher than acute STEMI patients. This dysfunction persists even up to 4 months after the acute event, whereas STEMI patients show recovery of autonomic function in this 4 month period. DC remained consistent between TTS and STEMI patients across all time points. Further research elucidating the long-term implications of these autonomic alterations is warranted to refine risk stratification and therapeutic strategies in TTS management.

Smartphone motion sensor based detection of atrial fibrillation in populations referred to cardiology

Abstract Introduction The motion sensors of a common smartphone when placed on chest can be used to detect vibrations caused by cardiac motion. The concept has been shown to detect atrial fibrillation (AFib) in clinical setting with 96% accuracy. Purpose We sought to examine how the smartphone motion sensor based AFib detection performs in various populations referred to cardiology pooled from 3 different studies addressing arrhythmias, heart failure or suspected coronary artery disease. Methods From a total of 1 524 patient recordings 748 recordings were from men (49%). The mean age of the patients was 64 years (range 20-94 years) and mean BMI was 28.9kg/m2 (range 15.2-63.7kg/m2). The signals were collected by placing a smartphone on chest of the patients in supine position and then analyzed with the CardioSignal solution. From the recordings two 60 second strips were analyzed. A simultaneous 5-lead or 1-lead ECG data was collected for reference. The ECG signals were analyzed for quality and rhythm diagnostics by 2 physicians blinded for the motion sensor AFib detection result. Finally, a head-to-head comparison was conducted for the motion sensor AFib detection result and the ECG data. Results From the 1524 patient recordings the final analysis set included 237 individuals with AFib, whereas the rest had sinus rhythm. The sensitivity to detect AFib was 91.9% and specificity 98.7%. The positive predictive value was 91.0% (88.3% to 94.2%) and the negative predictive value 98.3% (95.2% to 99.1%). The accuracy was 97.0% (94.7% to 97.7%). Conclusions A smartphone-based assessment of AFib using the embedded motion sensors seem to yield very good detection results, especially specificity. The accuracy of the results is comparable to available ECG and PPG based AFib detection solutions.

The added value of deep learning to submaximal exercise electrocardiogram for the 10-year prediction of major adverse cardiovascular and cerebrovascular events

Abstract Background The added value of exercise ECG data to traditional risk factors in predicting cardiovascular events remains unclear. Deep learning of ECG signals has demonstrated state-of-the-art performance in detecting subtle abnormalities indicative of cardiovascular disease. We investigated whether deep learning analysis of exercise ECGs could improve the prediction of major cardiovascular and cerebrovascular events (MACCE) with respect to established risk models in a large population-based cohort. Purpose To assess the added value of submaximal exercise ECG measurements to established cardiovascular risk prediction models. Methods We obtained ECG recordings from 41,076 UK Biobank participants without cardiovascular disease (CVD) who underwent a submaximal exercise ECG test. We obtained ECG recordings from 41,076 UK Biobank participants without cardiovascular disease (CVD) who underwent a submaximal exercise ECG test. We created two deep neural network ECG risk scores: one derived from conventional ECG parameters, measured at rest, peak exercise and late-stage recovery, and another based on the complete ECG (Q-ECG). We assessed the added value of conventional ECG parameters and Q-ECG risk scores to the UK's current prediction algorithm (QRISK3). We estimate the association between ECG parameters and MACCE, using Cox Proportional hazard models, following adjustment for traditional risk factors. All models were internally validated via 5-fold cross-validation and 1000 bootstrap iterations. Predictive performance was evaluated using Harrel's C-index, Net Reclassification Index (NRI) and net benefit. Findings: Incident MACCE was reported in 4,082 (9.9%) individuals in the study population and 3,463(9.7%) individuals with valid ECG parameters over a median follow-up period of 12.5 years. We found combined conventional ECG and Q-ECG scores were independently associated with MACCE, following adjustment for multiple testing: adjusted hazard ratio [HZ] = 1.76 (95% CI:1.63-1.91); HZ = 1.14 (95% CI:1.10-1.18), respectively, per standard deviation increase. Both conventional ECG parameters and Q-ECG were predictive of MACCE, independent of clinical risk factors, C-index = 0.64 (95%CI 0.63-0.65); net benefit = 0.09(95% CI 0.07 - 0.11) and C-index = 0.56 (95% CI 0.55-0.57); net benefit = 0.07(95% CI 0.05 - 0.09). ECG measurement's modestly improved model discrimination over the bassline QRISK3 risk score when combined with QRISK3 risk factors for conventional markers and Q-ECG score, respectively; ΔC-index 0.03 (95% CI: 0.02 – 0.04) and ΔC-index 0.03 (95% CI: 0.02 – 0.03); However, we observed no significant improvements in classification at the current recommended threshold of 10%. Conclusion In individuals without a history of prior cardiovascular disease, ECG measures independently predict the risk of MACCE. When combined with QRISK3, neural-network-derived ECG risk scores marginally improve cardiovascular risk prediction over QRISK3.

Analysis of Physiological Signals to Detect the Symptoms of Paralysis Disorder using Neural Network Approach: a Cohort Study

As per World Medical Associations, Paralysis attack is one of the leading causes of disorder in the human subjects after other chronic disorders such as Cancer and Heart Attack. Due to the attack of Paralysis, the body parts of human body become un responsive due to which the person may be physically disabled from hands or legs-based locomotion. There is scarcity in the methodologies who can early indicate the symptoms and alert the medical practitioners or human subject under surveillance. In this study, the health-related signals were studied and as per consultation of medical experts, total eight key parameters were identified and then based on that the output decision that weather person has symptoms of paralysis or not. The observatory data base has been prepared from more than 40,000 human subjects for the duration January 2018 to December 2023. As per categorical analysis based on training, validation and testing of three-layer Recurrent Neural Network, it has been observed that the ECG, Blood pressure, Heart Rate and PPG signals are key indicators to judge the susceptance of Paralysis disorder in the human subjects. Based on the evaluation parameters of machine learning models, the conjunction of response in association with selected input parameters gives 94% accuracy than other models. Our study helps research community to develop such methodologies by which the early indication of such disorders can be helpful to human health.

Advanced Anomaly Detection in ECG Signals Through Convolutional Autoencoders

This article aims to present a comprehensive study on convolutional autoencoders for advanced anomaly detection in ECG signals. Anomaly detection in complex datasets has become increasingly critical due to the rising need for systems that can effectively identify irregularities that may indicate fraud, system failures, or significant deviations from normal operations. Traditional methods often need help capturing nuanced patterns in high-dimensional data, necessitating more sophisticated approaches. This research uses an autoencoder-based model as a robust solution for anomaly detection, utilizing its capability to learn high-level representations in an unsupervised manner. The proposed model uses a convolutional autoencoder architecture to compress and decompress input data, thus highlighting anomalies through reconstruction errors. We outline detailed experiment strategies, including model training on average data to minimize reconstruction loss, setting an optimal threshold for anomaly sensitivity based on validation loss, and evaluating the model using precision, recall, F1-score, and AUC-ROC metrics. These experiments were conducted using a dataset with labeled normal and abnormal instances, allowing precise tuning and assessment of model performance. The results indicate that the autoencoder discriminates between normal and abnormal data, achieving high precision and recall at 99.22% and 98.98%, respectively. The confusion matrix and loss distribution analysis further validate the model's efficacy, clearly distinguishing between normal and abnormal data loss values concerning the defined threshold. This research shows the autoencoder model demonstrates high accuracy in anomaly detection and offers insights into the types of anomalies it can detect, supporting its application across various domains requiring reliable anomaly identification.

RU-Net, Multi-Scale Feature Fusion and Transfer Learning: Unlocking the Potential of Cuffless Blood Pressure Monitoring with PPG and ECG.

This study introduces an innovative deep-learning model for cuffless blood pressure estimation using PPG and ECG signals, demonstrating state-of-the-art performance on the largest clean dataset, PulseDB. The rU-Net architecture, a fusion of U-Net and ResNet, enhances both generalization and feature extraction accuracy. Accurate multi-scale feature capture is facilitated by short-time Fourier transform (STFT) time-frequency distributions and multi-head attention mechanisms, allowing data-driven feature selection. The inclusion of demographic parameters as supervisory information further elevates performance. On the calibration-based dataset, our model excels, achieving outstanding accuracy (SBP MAE ± std: 4.49 ± 4.86 mmHg, DBP MAE ± std: 2.69 ± 3.10 mmHg), surpassing AAMI standards and earning a BHS Grade A rating. Addressing the challenge of calibration-free data, we propose a fine-tuning-based transfer learning approach. Remarkably, with only 10% data transfer, our model attains exceptional accuracy (SBP MAE ± std: 4.14 ± 5.01 mmHg, DBP MAE ± std: 2.48 ± 2.93 mmHg). This study sets the stage for the development of highly accurate and reliable wearable cuffless blood pressure monitoring devices.