Respiratory Rate Monitoring Research Articles

Validation of a Wearable System for Respiratory Rate Monitoring in Dogs

Validation of a Wearable System for Respiratory Rate Monitoring in Dogs

Video-Based Respiratory Rate Estimation for Infants in the NICU.

Non-contact respiratory rate estimation (RR) is highly desirable for infants because of their sensitive skin. We propose a novel RGB video-based RR estimation method for infants in the neonatal intensive care unit (NICU) that can accurately measure the RR contact-less. We utilize Eulerian video magnification (EVM) method and develop an adaptive peak prominence threshold value estimation method to address challenges of RR estimation (e.g., dark environments, shallow breathing, babies swaddled or under blankets). We recruited 13 infants recorded for 4 consecutive hours per case. We then evaluate the performance of the algorithm for several (i.e., 19 to 25) randomly selected videos, each lasting 1 minute, for each case. Intraclass correlation coefficients of the proposed method over manually and automatically selected ROIs are 0.91 (95%CI: [Formula: see text]) and 0.88 (95%CI: [Formula: see text]), indicating excellent and good reliability, respectively. The Bland-Altman analysis of the proposed algorithm shows higher agreement between the estimated values via the proposed method and visually counted RR than the agreement between the RR obtained from the impedance sensors and reference RR, and agreement between a former EVM-based method and reference RR values. Our algorithm shows promising results for RR estimation in a real-life NICU environment under various conditions that can confound the estimation. We present a robust algorithm for non-contact neonatal respiratory rate monitoring, capable of performing well under various environmental lighting conditions in NICU, even when the infant is clothed or covered.

Screen-printed wearable sensors for continuous respiratory rate monitoring: fabrication, clinical evaluation, and point-of-care potential

A screen-printed wearable sensor accurately monitored respiratory rate in 15 participants with a 0.055 bpm error, showing promise for point-of-care use in conditions like COPD and COVID-19.

An accurate respiratory rate estimation algorithm for a rubber fiber respiratory sensor

Respiratory measurement is a crucial indicator for assessing health status; however, current methods for measuring respiratory rate and frequency are passive and not intuitive. This study investigates peak detection algorithms using a resistive strain sensor integrated into a garment for respiratory rate monitoring. The sensor, constructed with CNT material and a flexible rubber substrate, exhibits high conformity to the body’s contours. Designed for respiration measurement, the sensor maintains a low 6% strain for optimal sensitivity, demonstrating a 4% decrease after 800 repetitions of 10% elongation. Garment design emphasizes cohesion between the sensor and fabric, achieved through a piping technique. Respiratory measurement relies on a resistive sensor principle, where abdominal volume changes induce tension, altering resistance. Three peak detection algorithms are evaluated: the window size algorithm, low-pass filter, and FIR filter. The window size algorithm shows a 93% matching rate for normal breathing but requires manual adjustments based on breathing speed. The low-pass filter reduces noise but introduces lag, challenging peak matching. The FIR filter effectively detects peaks at increased speeds, achieving a matching rate exceeding 98%. The study concludes that the choice of algorithm depends on respiratory scenarios, with the window size algorithm suitable for regular cycles, the low-pass filter for real-time monitoring, and the FIR filter for accelerated respiratory rates. The study primarily explores static situations, indicating the need for future research on dynamic respiratory movements to enhance algorithm versatility.

Comparison of the incidence of severe outcomes in outpatients with COVID-19 or seasonal influenza without risk factors: Retrospective analysis of a health insurance claims-database

BackgroundTo evaluate disease severity for coronavirus disease of 2019 (COVID-19), seasonal influenza (FLU) is often used as a control group for comparison. However, evidence regarding outpatients without risk factors for severe disease is insufficient to consider future treatment policies for these patients. ObjectivesTo compare the risk of severe disease in patients without risk factors for severe disease who are diagnosed with COVID-19 or FLU. Study designThis was a retrospective study analyzing a health insurance claims database (data collection: COVID-19, January 2020 through August 2022; FLU, October 2017 through September 2019). Outcome measures were incidences of severe events of hospitalization, respiratory and heart rate monitoring, oxygen therapy, ventilator use, and intensive care unit admission (within 28 days from diagnosis), and death (within the following month). ResultsData from 575,293 patients were included in the COVID-19 group and from 1,095,698 patients in the FLU group, without risk factors. Adjusted odds ratios for all study outcomes were significantly higher for patients with COVID-19 than with FLU (>3 times greater odds ratio for all severe events). For the population aged ≥18 years, there was also a significantly higher risk for all severe events with COVID-19 versus FLU. The incidence of severe events with COVID-19 among the population aged <18 years was greatly reduced (versus ≥18 years), and risk for severe events was not significantly different with COVID-19 versus FLU. ConclusionsThese findings suggest that the medical need for specific treatment for patients with COVID-19 may be as high as for FLU.

JoulesEye

Smartphones and smartwatches have contributed significantly to fitness monitoring by providing real-time statistics, thanks to accurate tracking of physiological indices such as heart rate. However, the estimation of calories burned during exercise is inaccurate and cannot be used for medical diagnosis. In this work, we present JoulesEye, a smartphone thermal camera-based system that can accurately estimate calorie burn by monitoring respiration rate. We evaluated JoulesEye on 54 participants who performed high intensity cycling and running. The mean absolute percentage error (MAPE) of JoulesEye was 5.8%, which is significantly better than the MAPE of 37.6% observed with commercial smartwatch-based methods that only use heart rate. Finally, we show that an ultra-low-resolution thermal camera that is small enough to fit inside a watch or other wearables is sufficient for accurate calorie burn estimation. These results suggest that JoulesEye is a promising new method for accurate and reliable calorie burn estimation.

Efficacy of Marker-Based Motion Capture for Respiratory Cycle Measurement: A Comparison with Spirometry.

Respiratory rate monitoring is fundamental in clinical settings, and the accuracy of measurement methods is critical. This study aimed to develop and validate methods for assessing respiratory rate and the duration leof respiratory cycle phases in different body positions using optoelectronic plethysmography (OEP) based on a motion capture video system. Two analysis methods, the summation method and the triangle method were developed. The study focused on determining the optimal number of markers while achieving accuracy in respiratory parameter measurements. The results showed that most analysis methods showed a difference of ≤0.5 breaths per minute, with R2 ≥ 0.94 (p < 0.001) compared to spirometry. The best OEP methods for respiratory rate were the abdominal triangles and the sum of abdominal markers in all body positions. The study explored inspiratory and expiratory durations. The research found that 5-9 markers were sufficient to accurately determine respiratory time components in all body positions, reducing the marker requirements compared to previous studies. This interchangeability of OEP methods with standard spirometry demonstrates the potential of non-invasive methods for the simultaneous assessment of body segment movements, center of pressure dynamics, and respiratory movements. Future research is required to improve the clinical applicability of these methods.

Stretchable polymer optical fiber embedded in the mattress for respiratory and heart rate monitoring

In this study, a stretchable polymer optical fiber with a sandwich structure was proposed, which integrated a silicone tube filled with polydimethylsiloxane (PDMS) with two common commercial polymethylmethacrylate (PMMA) optical fiber. By integrating the stretchable polymer optical fiber into the designed sleep-monitoring mattress, the pressure sensitivity of the mattress to measure human respiratory rate, heart rate, and body motion information is improved. The weak respiratory and heartbeat signals were amplified by an amplifier circuit and the respiratory and heart rates were obtained by an algorithm. The experimental results indicated that the mattress can monitor respiratory rate and heart rate values in different positions as well as body movement information and can differentiate between three different behavioral states, including not on the bed, lying, and moving on the bed. In addition, clinical trials were conducted on 10 subjects, and the results were analyzed using the Bland-Altman analysis method, confirming the accuracy of the mattress, with maximum errors of less than 1 breath per minute and 2 beats per minute for respiration and heartbeat respectively. In addition, the mattress's reliability and stability were verified through continuous sleep monitoring at various times of day. The mattress has the advantages of low cost, softness, comfort, and high sensitivity, and is suitable as a home medical device for long-term continuous monitoring of respiratory and heart rates during sleep.

Accurate and continuous respiratory rate using touchless monitoring technology

PurposeRespiratory rate is a commonly used vital sign with various clinical applications. It serves as a crucial marker of acute health issues and any significant alteration in respiratory rate may be an early warning sign of major issues such as infections in the respiratory tract, respiratory failure, or cardiac arrest. Timely recognition of changes in respiratory rate enables prompt medical action, while neglecting to detect a change may lead to adverse patient outcomes. Here, we report on the performance of respiratory rate determined using a depth sensing camera system (RRdepth) which allows for continuous, non-contact ‘touchless’ monitoring of this important vital sign. MethodsThirty adult volunteers undertook a range of set breathing rates to cover a target breathing range of 4–40 breaths/min. Depth information was acquired from the torso region of the subjects using an Intel D415 RealSense camera positioned above the bed. The depth information was processed to generate a respiratory signal from which RRdepth was calculated. This was compared to a manually scored capnograph reference (RRcap). ResultsAn overall RMSD accuracy of 0.77 breaths/min was achieved across the target respiratory rate range with a corresponding bias of 0.05 breaths/min. This corresponded to a line of best fit given by RRdepth = 1.01 x RRcap – 0.22 breaths/min with an associated high degree of correlation (R = 0.997). A breakdown of the performance with respect to sub-ranges corresponding to respiratory rates or ≤7, >7-10, >10-20, >20-30, >30 breaths/min all exhibited RMSD accuracies of less than 1.00 breaths/min. We also had the opportunity to test the performance of spontaneous breathing of the subjects which occurred during the study and found an overall RMSD accuracy of 1.20 breaths/min with corresponding accuracies ≤1.30 breaths/min across each of the individual sub-ranges. ConclusionsWe have conducted an investigative study of a prototype depth sensing camera system for the non-contact monitoring of respiratory rate. The system achieved good performance with high accuracy across a wide range of rates including both clinically important high and low rates.

Home-based initiatives for acute management of COVID-19 patients needing oxygen: differences across The Netherlands

ObjectiveDuring the COVID-19 pandemic new collaborative-care initiatives were developed for treating and monitoring COVID-19 patients with oxygen at home. Aim was to provide a structured overview focused on differences and similarities of initiatives of acute home-based management in the Netherlands.MethodsInitiatives were eligible for evaluation if (i) COVID-19 patients received oxygen treatment at home; (ii) patients received structured remote monitoring; (iii) it was not an ‘early hospital discharge’ program; (iv) at least one patient was included. Protocols were screened, and additional information was obtained from involved physicians. Design choices were categorised into: eligible patient group, organization medical care, remote monitoring, nursing care, and devices used.ResultsNine initiatives were screened for eligibility; five were included. Three initiatives included low-risk patients and two were designed specifically for frail patients. Emergency department (ED) visit for an initial diagnostic work-up and evaluation was mandatory in three initiatives before starting home management. Medical responsibility was either assigned to the general practitioner or hospital specialist, most often pulmonologist or internist. Pulse-oximetry was used in all initiatives, with additional monitoring of heart rate and respiratory rate in three initiatives. Remote monitoring staff’s qualification and authority varied, and organization and logistics were covered by persons with various backgrounds. All initiatives offered remote monitoring via an application, two also offered a paper diary option.ConclusionsWe observed differences in the organization of interprofessional collaboration for acute home management of hypoxemic COVID-19 patients. All initiatives used pulse-oximetry and an app for remote monitoring. Our overview may be of help to healthcare providers and organizations to set up and implement similar acute home management initiatives for critical episodes of COVID-19 (or other acute disorders) that would otherwise require hospital care.

Use of a wearable patch sensor for continuous activity and posture tracking to assess risk of early infection after transcatheter aortic valve replacement

Abstract Background Early infections after transcatheter aortic valve replacement (TAVR) represent one of the most common risk factors for non-cardiac 30-day mortality. Several wearable devices have been developed to measure vital signs and physical activity in real time while allowing full mobilization of the patient. However, their use has not been evaluated in a clinical setting with patients undergoing TAVR yet. Purpose We aimed to evaluate the feasibility of wearable patch sensors in assessing vital signs and daily activity before and after TAVR to predict an increased risk of early infections following TAVR. Methods Patients at high risk for infection undergoing transfemoral TAVR under local anesthesia were enrolled between September 2022 and January 2023. Postprocedural early infections were assessed, defined as C-reactive protein rise within 5 days after the procedure inducing antibiotic treatment. Vital signs (heart rate, respiratory rate, single-lead ECG) and activity (steps, posture) were monitored by a wearable patch sensor two days before (pre-TAVR) and two days after TAVR (post-TAVR). Results 54 patients completed 5 days of in-hospital monitoring with wearable patch sensors. The mean age was 80.3+5.2 years, 31 patients (57.4%) were male. Early infections occurred in 8 of 54 (14.8%) high-risk patients, of whom one was a urinary tract, one a pulmonary, and two access site-related etiology. Four infections remained with unknown focus. Respiratory rate was generally higher in patients developing an early infection after TAVR in both time periods, before (early infection 22±3 vs. no infection 18±2 bpm; p=0.001) and after TAVR (early infection 24±3 vs. no infection 19±3 bpm; p&amp;lt;0.001). Early infections were more prevalent in patients who showed less activity in the two days before TAVR regarding lower step count (early infection 242±360 vs. no infection 515±566; p=0.048) and high percentage of lying position during the whole day (early infection 89±12% vs. no infection 64±23%; p=0.071). Two days after TAVR, there were no significant differences in activity levels between both cohorts. Physical activity was reduced in both groups compared to the two days before TAVR, resulting in a reduced number of steps (pre-TAVR 473±546 vs. post-TAVR 267±434; p=0.039) and more daytime spent in lying position (pre-TAVR 57±19% vs. post-TAVR 73±24%; p&amp;lt;0.0001). Conclusion Wearable patch sensors are feasible to track vital signs and physical activity during hospital stay with promising opportunity for improving clinical management in TAVR patients. Early infections after TAVR are associated with decreased preprocedural mobility in high-risk patients. Therefore, these patients may be prematurely detected using wireless respiratory rate monitoring and infections prevented by intensified physiotherapy.

Graphene/polypyrrole/carbon black nanocomposite material ink-based screen-printed low-cost, flexible humidity sensor

The detection of humidity is crucial for various applications in industry, healthcare, and the environment. To meet the needs of many of these applications, humidity sensors must be flexible, disposable, and easily fabricated. This article introduces a cost-effective and flexible humidity sensor created on a paper substrate, using graphene-polypyrrole-carbon black ink. The sensor shows excellent sensing capabilities, with a resistance change of approximately 12.2 Ω/%RH when exposed to humidity ranging from 23%RH to 92.7%RH. Additionally, the sensor is highly flexible, stable, and repeatable for over 50 cycles, with a short response/recovery time of approximately 5 s/7 s for respiration rate monitoring. Furthermore, the sensor demonstrates good reproducibility, with minor variations of approximately ± 1 Ω/%RH. The performance of the produced humidity sensor is assessed for monitoring humidity in a spatial setting as well as monitoring soil moisture. As a result, these findings indicate that the proposed humidity sensor is resilient enough to be used in wearable and flexible electronic devices.

Early detection of deterioration in COVID-19 patients by continuous ward respiratory rate monitoring: a pilot prospective cohort study.

Tachypnea is among the earliest signs of pulmonary decompensation. Contactless continuous respiratory rate monitoring might be useful in isolated COVID-19 patients admitted in wards. We therefore aimed to determine whether continuous monitoring of respiratory patterns in hospitalized patients with COVID-19 predicts subsequent need for increased respiratory support. Single-center pilot prospective cohort study in COVID-19 patients who were cared for in routine wards. COVID-19 patients who had at least one escalation of pulmonary management were matched to three non-escalated patients. Contactless respiratory monitoring was instituted after patients enrolled, and continued for 15 days unless hospital discharge, initiation of invasive mechanical ventilation, or death occurred. Clinicians were blinded to respiratory rate data from the continuous monitor. The exposures were respiratory features over rolling periods of 30 min, 24 h, and 72 h before respiratory care escalation. The primary outcome was a subsequent escalation in ventilatory support beyond a Venturi mask. Among 125 included patients, 13 exhibited at least one escalation and were each matched to three non-escalated patients. A total of 28 escalation events were matched to 84 non-escalation episodes. The 30-min mean respiratory rate in escalated patients was 23 breaths per minute (bpm) ranging from 13 to 40 bpm, similar to the 22 bpm in non-escalated patients, although with less variability (range 14 to 31 bpm). However, higher respiratory rate variability, especially skewness over 1 day, was associated with higher incidence of escalation events. Our overall model, based on continuous data, had a moderate accuracy with an AUC 0.81 (95%CI: 0.73, 0.88) and a good specificity 0.93 (95%CI: 0.87, 0.99). Our pilot observational study suggests that respiratory rate variability as detected with continuous monitoring is associated with subsequent care escalation during the following 24 h. Continuous respiratory monitoring thus appears to be a valuable increment over intermittent monitoring. Our study was the initial evaluation of Circadia contactless respiratory monitoring in COVID-19 patients who are at special risk of pulmonary deterioration. The major limitation is that the analysis was largely post hoc and thus needs to be confirmed in an out-of-sample population.

Nylon Fabric/GO Based Self-Powered Humidity Sensor Based on the Galvanic Cell Principle with High Air Permeability and Rapid-Response.

Flexible humidity sensors have received more and more attention in people's lives, and the problems of gas permeability and power supply issues of the device have long been areas in need of improvement. In this work, inspired by the high air permeability of daily wear clothing and galvanic batteries, a self-powered humidity sensor with high air permeability and fast response is designed. A nylon fabric/GO net (as a humidity sensitive layer and solid electrolyte) is obtained by spraying technique. This structure enables the sensor to have fast response/recovery (0.78s/0.93s, calculated at 90% of the final value), ultra-high response (0.83V) and excellent stability (over 150 cycles) at 35°C. Such sensors are useful for health monitoring, such as non-contact monitoring of human respiratory rate before and after exercise, and monitoring a level of humidity in the palms, arms, and fingers. This research provides an idea for developing a flexible wearable humidity sensor that is both breathable and self-powered and can also be mass-produced similar to wearable clothing.

Cardiorespiratory monitoring with a wireless and nonadhesive belt measuring diaphragm activity in preterm and term infants: A multicenter non-inferiority study.

We determined if the heart rate (HR) monitoring performance of a wireless and nonadhesive belt is non-inferior compared to standard electrocardiography (ECG). Secondary objective was to explore the belt's respiratory rate (RR) monitoring performance compared to chest impedance (CI). In this multicenter non-inferiority trial, preterm and term infants were simultaneously monitored with the belt and conventional ECG/CI for 24 h. HR monitoring performance was estimated with the HR difference and ability to detect cardiac events compared to the ECG, and the incidence of HR-data loss per second. These estimations were statistically compared to prespecified margins to confirm equivalence/non-inferiority. Exploratory RR analyses estimated the RR trend difference and ability to detect apnea/tachypnea compared to CI, and the incidence of RR-data loss per second. Thirty-nine infants were included. HR monitoring with the belt was non-inferior to the ECG with a mean HR difference of 0.03 beats per minute (bpm) (standard error [SE] = 0.02) (95% limits of agreement [LoA]: [-5 to 5] bpm) (p < 0.001). Second, sensitivity and positive predictive value (PPV) for cardiac event detection were 94.0% (SE = 0.5%) and 92.6% (SE = 0.6%), respectively (p ≤ 0.001). Third, the incidence of HR-data loss was 2.1% (SE = 0.4%) per second (p < 0.05). The exploratory analyses of RR showed moderate trend agreement with a mean RR-difference of 3.7 breaths/min (SE = 0.8) (LoA: [-12 to 19] breaths/min), but low sensitivities and PPV's for apnea/tachypnea detection. The incidence of RR-data loss was 2.2% (SE = 0.4%) per second. The nonadhesive, wireless belt showed non-inferior HR monitoring and a moderate agreement in RR trend compared to ECG/CI. Future research on apnea/tachypnea detection is required.

Contact methods for registering respiratory rate: opportunities and perspectives

Introduction. Respiratory rate is known to be one of the most important indicators reflecting the vital functions of a person. An increase in respiratory rate can be found in many diseases and pathological conditions, for example, in chronic obstructive pulmonary disease, pneumonia, bronchial asthma, myocardial infarction, heart failure, anaemia, etc. Due to the active introduction of telemedicine monitoring into clinical practice, the measurement of the abovementioned indicator is particularly relevant for the purpose of early detection and prevention of complications of chronic non-infectious diseases, as well as dynamic monitoring of the condition of patients in both inpatient and outpatient settings.Aim. To search and update information about existing and promising developments for the control of respiratory rate based on different physical principles.Materials and methods. For this review we used databases PubMed, Scopus, MedLine and eLIBRARY. The following keywords were used for the search: “respiratory rate”, “contact”, “measurement”, “sensor”.Results. Contact methods for measuring respiratory rate include a wide range of sensors based on various physical principles. All types of sensors have their own application, but also they have some drawbacks. In order to achieve maximum accuracy of respiratory rate monitoring, it is necessary to carefully assess the conditions in which the patient is located, selecting the most appropriate technological solutions for them. Probably, complex systems, including several different sensors, are able to overcome many shortcomings. In addition, the development of information analysis methods, machine learning and artificial intelligence technologies can increase the sensitivity and accuracy of methods of measuring respiratory rate, reducing the frequency of bias associated with various artefacts.Conclusion. Thus, technological development opens up wide opportunities for long-term monitoring of vital functions, prevention and timely response to adverse events.

Effect of averaging time and respiratory pause time on the measurement of acoustic respiration rate monitoring

BackgroundAcoustic respiration rate (RRa) monitoring is a method of continuously measuring respiratory rate using a signal from an acoustic transducer placed over the airway. The purpose of the present study is to examine how the averaging time and respiratory pause time settings of an RRa monitor affect the detection time of sudden respiratory rate changes.MethodsA total of 40 healthy adult volunteers were included in the study. First, we measured the apnea detection time (apnea test) by dividing them into two groups (N = 20 each), one with a respiratory pause time setting of 20 s and the other with 40 s. Each group performed two apnea tests with an averaging time setting of 10 and 30 s. Next, we measured the tachypnea detection time (tachypnea test) for half of the subjects (N = 20) with two averaging time settings of 10 and 30 s. For each test, three measurements were taken, and the average of the three measurements was recorded.ResultsThere was no significant difference in the apnea detection time between the averaging time set at 10 and 30 s regardless of whether the respiratory pause time was set at 20 or 40 s. However, the apnea detection time was significantly shorter with the respiratory pause time of 20 s than 40 s, regardless of whether the averaging time was set at 10 or 30 s (p < 0.001). The tachypnea detection time was shorter with the averaging time of 10 s than 30 s (p < 0.001).Furthermore, the apnea detection time and tachypnea detection time were much longer than the actual settings.ConclusionsThe results of the current study show that in the measurement of RRa, the apnea detection time is more affected by the respiratory pause time setting than the averaging time setting; however, the tachypnea detection time is significantly affected by the averaging time setting.

Polyaniline/Ti3C2Tx functionalized mask sensors for monitoring of CO2 and human respiration rate

Flexible and wearable gas-sensing devices have become increasingly crucial for monitoring air quality and human health. Disposable masks, mainly composed of polypropylene (PP) fibers, can serve as a potential flexible substrate for sensing devices yet have been largely overlooked during the COVID-19 pandemic. In the present work, flexible Ti3C2Tx/PANI-PP composite gas sensors were fabricated by spray-coating delaminated Ti3C2Tx MXene and in-situ polymerization of aniline on a disposable mask substrate. The resulting hybrid gas sensor displayed a wide detection range (25–1500 ppm), reliable reproducibility, long-term stability, and excellent flexibility and selectivity, a remarkable response (15.2%) towards 500 ppm CO2 gas, which is 6.5 times higher than pristine Ti3C2Tx and 2.4 times higher than pristine PANI. The enhanced sensing performance of the composite sensor is mainly attributed to the synergistic effects of the heterojunctions between Ti3C2Tx and PANI, the improved conductivity, and the enlarged specific surface area. We further investigated the influence of various parameters, including humidity, temperature, bending angles, and folding times on the sensing performance of the composite gas sensor. Finally, a wearable wireless Bluetooth sensing device was fabricated for human exhaled breath monitoring at room temperature, demonstrating the potential application of this Ti3C2Tx/PANI-PP composite sensor for respiratory disease diagnosis.

RRWaveNet: A Compact End-to-End Multiscale Residual CNN for Robust PPG Respiratory Rate Estimation

Respiratory rate (RR) is an important biomarker as RR changes can reflect severe medical events such as heart disease, lung disease, and sleep disorders. Unfortunately, standard manual RR counting is prone to human error and cannot be performed continuously. This study proposes a method for continuously estimating RR, RRWaveNet. The method is a compact end-to-end deep learning model which does not require feature engineering and can use low-cost raw photoplethysmography (PPG) as input signal. RRWaveNet was tested subject-independently and compared to baseline in four datasets (BIDMC, CapnoBase, WESAD, and SensAI) and using three window sizes (16, 32, and 64 seconds). RRWaveNet outperformed current state-of-the-art methods with mean absolute errors at optimal window size of 1.66 1 1.01, 1.59 1 1.08, 1.92 1 0.96, and 1.23 1 0.61 breaths per minute for each dataset. In remote monitoring settings, such as in the WESAD and SensAI datasets, we apply transfer learning to improve the performance using two other ICU datasets as pretraining datasets, reducing the MAE by up to 21%. This shows that this model allows accurate and practical estimation of RR on affordable and wearable devices. Our study also shows feasibility of remote RR monitoring in the context of telemedicine and at home.

Ultra-Wideband Radar for Simultaneous and Unobtrusive Monitoring of Respiratory and Heart Rates in Early Childhood: A Deep Transfer Learning Approach.

Unobtrusive monitoring of children's heart rate (HR) and respiratory rate (RR) can be valuable for promoting the early detection of potential health issues, improving communication with healthcare providers and reducing unnecessary hospital visits. A promising solution for wireless vital sign monitoring is radar technology. This paper presents a novel approach for the simultaneous estimation of children's RR and HR utilizing ultra-wideband (UWB) radar using a deep transfer learning algorithm in a cohort of 55 children. The HR and RR are calculated by processing radar signals via spectrogram from time epochs of 10 s (25 sample length of hamming window with 90% overlap) and then transforming the resultant representation into 2-dimensional images. These images were fed into a pre-trained Visual Geometry Group-16 (VGG-16) model (trained on ImageNet dataset), with weights of five added layers fine-tuned using the proposed data. The prediction on the test data achieved a mean absolute error (MAE) of 7.3 beats per minute (BPM < 6.5% of average HR) and 2.63 breaths per minute (BPM < 7% of average RR). We also achieved a significant Pearson's correlation of 77% and 81% between true and extracted for HR and RR, respectively. HR and RR samples are extracted every 10 s.