1
|
Rathore KS, Vijayarangan S, Sp P, Sivaprakasam M. A Multifunctional Network with Uncertainty Estimation and Attention-Based Knowledge Distillation to Address Practical Challenges in Respiration Rate Estimation. SENSORS (BASEL, SWITZERLAND) 2023; 23:s23031599. [PMID: 36772640 PMCID: PMC9920118 DOI: 10.3390/s23031599] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 12/15/2022] [Revised: 01/20/2023] [Accepted: 01/23/2023] [Indexed: 05/03/2023]
Abstract
Respiration rate is a vital parameter to indicate good health, wellbeing, and performance. As the estimation through classical measurement modes are limited only to rest or during slow movements, respiration rate is commonly estimated through physiological signals such as electrocardiogram and photoplethysmography due to the unobtrusive nature of wearable devices. Deep learning methodologies have gained much traction in the recent past to enhance accuracy during activities involving a lot of movement. However, these methods pose challenges, including model interpretability, uncertainty estimation in the context of respiration rate estimation, and model compactness in terms of deployment in wearable platforms. In this direction, we propose a multifunctional framework, which includes the combination of an attention mechanism, an uncertainty estimation functionality, and a knowledge distillation framework. We evaluated the performance of our framework on two datasets containing ambulatory movement. The attention mechanism visually and quantitatively improved instantaneous respiration rate estimation. Using Monte Carlo dropouts to embed the network with inferential uncertainty estimation resulted in the rejection of 3.7% of windows with high uncertainty, which consequently resulted in an overall reduction of 7.99% in the mean absolute error. The attention-aware knowledge distillation mechanism reduced the model's parameter count and inference time by 49.5% and 38.09%, respectively, without any increase in error rates. Through experimentation, ablation, and visualization, we demonstrated the efficacy of the proposed framework in addressing practical challenges, thus taking a step towards deployment in wearable edge devices.
Collapse
Affiliation(s)
- Kapil Singh Rathore
- Indian Institute of Technology Madras, Chennai 6000001, India
- Healthcare Technology Innovation Center, Chennai 6000001, India
| | - Sricharan Vijayarangan
- Indian Institute of Technology Madras, Chennai 6000001, India
- Healthcare Technology Innovation Center, Chennai 6000001, India
| | - Preejith Sp
- Healthcare Technology Innovation Center, Chennai 6000001, India
| | - Mohanasankar Sivaprakasam
- Indian Institute of Technology Madras, Chennai 6000001, India
- Healthcare Technology Innovation Center, Chennai 6000001, India
| |
Collapse
|
2
|
Chen Y, Sun Y, Lv J, Jia B, Huang X. End-to-end heart sound segmentation using deep convolutional recurrent network. COMPLEX INTELL SYST 2021. [DOI: 10.1007/s40747-021-00325-w] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/16/2022]
Abstract
AbstractHeart sound segmentation (HSS) aims to detect the four stages (first sound, systole, second heart sound and diastole) from a heart cycle in a phonocardiogram (PCG), which is an essential step in automatic auscultation analysis. Traditional HSS methods need to manually extract the features before dealing with HSS tasks. These artificial features highly rely on extraction algorithms, which often result in poor performance due to the different operating environments. In addition, the high-dimension and frequency characteristics of audio also challenge the traditional methods in effectively addressing HSS tasks. This paper presents a novel end-to-end method based on convolutional long short-term memory (CLSTM), which directly uses audio recording as input to address HSS tasks. Particularly, the convolutional layers are designed to extract the meaningful features and perform the downsampling, and the LSTM layers are developed to conduct the sequence recognition. Both components collectively improve the robustness and adaptability in processing the HSS tasks. Furthermore, the proposed CLSTM algorithm is easily extended to other complex heart sound annotation tasks, as it does not need to extract the characteristics of corresponding tasks in advance. In addition, the proposed algorithm can also be regarded as a powerful feature extraction tool, which can be integrated into the existing models for HSS. Experimental results on real-world PCG datasets, through comparisons to peer competitors, demonstrate the outstanding performance of the proposed algorithm.
Collapse
|
3
|
Sports medicine: bespoke player management. Digit Health 2021. [DOI: 10.1016/b978-0-12-818914-6.00021-1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/19/2022] Open
|
4
|
Chen Y, Lv J, Sun Y, Jia B. Heart sound segmentation via Duration Long–Short Term Memory neural network. Appl Soft Comput 2020. [DOI: 10.1016/j.asoc.2020.106540] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/23/2022]
|
5
|
Liu H, Allen J, Zheng D, Chen F. Recent development of respiratory rate measurement technologies. Physiol Meas 2019; 40:07TR01. [PMID: 31195383 DOI: 10.1088/1361-6579/ab299e] [Citation(s) in RCA: 93] [Impact Index Per Article: 18.6] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/12/2022]
Abstract
Respiratory rate (RR) is an important physiological parameter whose abnormality has been regarded as an important indicator of serious illness. In order to make RR monitoring simple to perform, reliable and accurate, many different methods have been proposed for such automatic monitoring. According to the theory of respiratory rate extraction, methods are categorized into three modalities: extracting RR from other physiological signals, RR measurement based on respiratory movements, and RR measurement based on airflow. The merits and limitations of each method are highlighted and discussed. In addition, current works are summarized to suggest key directions for the development of future RR monitoring methodologies.
Collapse
Affiliation(s)
- Haipeng Liu
- Faculty of Health, Education, Medicine, and Social Care, Anglia Ruskin University, Chelmsford, CM1 1SQ, United Kingdom. Department of Electrical and Electronic Engineering, Southern University of Science and Technology, Shenzhen 518055, People's Republic of China
| | | | | | | |
Collapse
|
6
|
Janbakhshi P, Shamsollahi MB. ECG-derived respiration estimation from single-lead ECG using gaussian process and phase space reconstruction methods. Biomed Signal Process Control 2018. [DOI: 10.1016/j.bspc.2018.05.025] [Citation(s) in RCA: 17] [Impact Index Per Article: 2.8] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/26/2022]
|
7
|
Dias D, Paulo Silva Cunha J. Wearable Health Devices-Vital Sign Monitoring, Systems and Technologies. SENSORS (BASEL, SWITZERLAND) 2018; 18:E2414. [PMID: 30044415 PMCID: PMC6111409 DOI: 10.3390/s18082414] [Citation(s) in RCA: 239] [Impact Index Per Article: 39.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 05/31/2018] [Revised: 07/19/2018] [Accepted: 07/21/2018] [Indexed: 01/05/2023]
Abstract
Wearable Health Devices (WHDs) are increasingly helping people to better monitor their health status both at an activity/fitness level for self-health tracking and at a medical level providing more data to clinicians with a potential for earlier diagnostic and guidance of treatment. The technology revolution in the miniaturization of electronic devices is enabling to design more reliable and adaptable wearables, contributing for a world-wide change in the health monitoring approach. In this paper we review important aspects in the WHDs area, listing the state-of-the-art of wearable vital signs sensing technologies plus their system architectures and specifications. A focus on vital signs acquired by WHDs is made: first a discussion about the most important vital signs for health assessment using WHDs is presented and then for each vital sign a description is made concerning its origin and effect on heath, monitoring needs, acquisition methods and WHDs and recent scientific developments on the area (electrocardiogram, heart rate, blood pressure, respiration rate, blood oxygen saturation, blood glucose, skin perspiration, capnography, body temperature, motion evaluation, cardiac implantable devices and ambient parameters). A general WHDs system architecture is presented based on the state-of-the-art. After a global review of WHDs, we zoom in into cardiovascular WHDs, analysing commercial devices and their applicability versus quality, extending this subject to smart t-shirts for medical purposes. Furthermore we present a resumed evolution of these devices based on the prototypes developed along the years. Finally we discuss likely market trends and future challenges for the emerging WHDs area.
Collapse
Affiliation(s)
- Duarte Dias
- Biomedical Research and INnovation (BRAIN), Centre for Biomedical Engineering Research (C-BER), INESC Technology and Science, Porto 4200-465, Portugal.
| | - João Paulo Silva Cunha
- Biomedical Research and INnovation (BRAIN), Centre for Biomedical Engineering Research (C-BER), INESC Technology and Science, Porto 4200-465, Portugal.
- Faculty of Engineering, University of Porto, Porto 4200-465, Portugal.
| |
Collapse
|
8
|
Sharma H, Sharma KK. ECG-derived respiration based on iterated Hilbert transform and Hilbert vibration decomposition. AUSTRALASIAN PHYSICAL & ENGINEERING SCIENCES IN MEDICINE 2018; 41:429-443. [PMID: 29667117 DOI: 10.1007/s13246-018-0640-0] [Citation(s) in RCA: 9] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 08/16/2017] [Accepted: 04/11/2018] [Indexed: 11/26/2022]
Abstract
Monitoring of the respiration using the electrocardiogram (ECG) is desirable for the simultaneous study of cardiac activities and the respiration in the aspects of comfort, mobility, and cost of the healthcare system. This paper proposes a new approach for deriving the respiration from single-lead ECG based on the iterated Hilbert transform (IHT) and the Hilbert vibration decomposition (HVD). The ECG signal is first decomposed into the multicomponent sinusoidal signals using the IHT technique. Afterward, the lower order amplitude components obtained from the IHT are filtered using the HVD to extract the respiration information. Experiments are performed on the Fantasia and Apnea-ECG datasets. The performance of the proposed ECG-derived respiration (EDR) approach is compared with the existing techniques including the principal component analysis (PCA), R-peak amplitudes (RPA), respiratory sinus arrhythmia (RSA), slopes of the QRS complex, and R-wave angle. The proposed technique showed the higher median values of correlation (first and third quartile) for both the Fantasia and Apnea-ECG datasets as 0.699 (0.55, 0.82) and 0.57 (0.40, 0.73), respectively. Also, the proposed algorithm provided the lowest values of the mean absolute error and the average percentage error computed from the EDR and reference (recorded) respiration signals for both the Fantasia and Apnea-ECG datasets as 1.27 and 9.3%, and 1.35 and 10.2%, respectively. In the experiments performed over different age group subjects of the Fantasia dataset, the proposed algorithm provided effective results in the younger population but outperformed the existing techniques in the case of elderly subjects. The proposed EDR technique has the advantages over existing techniques in terms of the better agreement in the respiratory rates and specifically, it reduces the need for an extra step required for the detection of fiducial points in the ECG for the estimation of respiration which makes the process effective and less-complex. The above performance results obtained from two different datasets validate that the proposed approach can be used for monitoring of the respiration using single-lead ECG.
Collapse
Affiliation(s)
- Hemant Sharma
- Department of Electronics and Communication Engineering, National Institute of Technology Rourkela, Rourkela, India.
| | - K K Sharma
- Department of Electronics and Communication Engineering, Malaviya National Institute of Technology Jaipur, Jaipur, India
| |
Collapse
|
9
|
Ghaderyan P, Abbasi A. A novel cepstral-based technique for automatic cognitive load estimation. Biomed Signal Process Control 2018. [DOI: 10.1016/j.bspc.2017.07.020] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/19/2022]
|
10
|
Saatci E. Correlation analysis of respiratory signals by using parallel coordinate plots. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2018; 153:41-51. [PMID: 29157460 DOI: 10.1016/j.cmpb.2017.10.003] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 10/06/2016] [Revised: 05/08/2017] [Accepted: 10/02/2017] [Indexed: 06/07/2023]
Abstract
BACKGROUND AND OBJECTIVES The understanding of the bonds and the relationships between the respiratory signals, i.e. the airflow, the mouth pressure, the relative temperature and the relative humidity during breathing may provide the improvement on the measurement methods of respiratory mechanics and sensor designs or the exploration of the several possible applications in the analysis of respiratory disorders. Therefore, the main objective of this study was to propose a new combination of methods in order to determine the relationship between respiratory signals as a multidimensional data. METHODS In order to reveal the coupling between the processes two very different methods were used: the well-known statistical correlation analysis (i.e. Pearson's correlation and cross-correlation coefficient) and parallel coordinate plots (PCPs). Curve bundling with the number intersections for the correlation analysis, Least Mean Square Time Delay Estimator (LMS-TDE) for the point delay detection and visual metrics for the recognition of the visual structures were proposed and utilized in PCP. RESULTS The number of intersections was increased when the correlation coefficient changed from high positive to high negative correlation between the respiratory signals, especially if whole breath was processed. LMS-TDE coefficients plotted in PCP indicated well-matched point delay results to the findings in the correlation analysis. Visual inspection of PCB by visual metrics showed range, dispersions, entropy comparisons and linear and sinusoidal-like relationships between the respiratory signals. CONCLUSION It is demonstrated that the basic correlation analysis together with the parallel coordinate plots perceptually motivates the visual metrics in the display and thus can be considered as an aid to the user analysis by providing meaningful views of the data.
Collapse
Affiliation(s)
- Esra Saatci
- Department of Electrical and Electronics Engineering, Istanbul Kultur University, Bakirkoy, Istanbul, Turkey.
| |
Collapse
|
11
|
|
12
|
Charlton PH, Birrenkott DA, Bonnici T, Pimentel MAF, Johnson AEW, Alastruey J, Tarassenko L, Watkinson PJ, Beale R, Clifton DA. Breathing Rate Estimation From the Electrocardiogram and Photoplethysmogram: A Review. IEEE Rev Biomed Eng 2017; 11:2-20. [PMID: 29990026 PMCID: PMC7612521 DOI: 10.1109/rbme.2017.2763681] [Citation(s) in RCA: 127] [Impact Index Per Article: 18.1] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/11/2022]
Abstract
Breathing rate (BR) is a key physiological parameter used in a range of clinical settings. Despite its diagnostic and prognostic value, it is still widely measured by counting breaths manually. A plethora of algorithms have been proposed to estimate BR from the electrocardiogram (ECG) and pulse oximetry (photoplethysmogram, PPG) signals. These BR algorithms provide opportunity for automated, electronic, and unobtrusive measurement of BR in both healthcare and fitness monitoring. This paper presents a review of the literature on BR estimation from the ECG and PPG. First, the structure of BR algorithms and the mathematical techniques used at each stage are described. Second, the experimental methodologies that have been used to assess the performance of BR algorithms are reviewed, and a methodological framework for the assessment of BR algorithms is presented. Third, we outline the most pressing directions for future research, including the steps required to use BR algorithms in wearable sensors, remote video monitoring, and clinical practice.
Collapse
Affiliation(s)
- Peter H. Charlton
- Department of Biomedical Engineering, King’s College London, London SE1 7EH, U.K., and also with the Department of Engineering Science, University of Oxford, Oxford OX3 7DQ, U.K
| | - Drew A. Birrenkott
- Department of Engineering Science, University of Oxford, Oxford OX3 7DQ, U.K
| | - Timothy Bonnici
- Nuffield Department of Medicine, University of Oxford, Oxford OX3 9DU, U.K., and also with the Department of Asthma, Allergy, and Lung Biology, King’s College London, London SE1 7EH, U.K
| | | | - Alistair E. W. Johnson
- Laboratory for Computational Physiology, Massachusetts Institute of Technology, Cambridge, MA 02139 USA
| | - Jordi Alastruey
- Department of Biomedical Engineering, King’s College London, London SE1 7EH, U.K
| | - Lionel Tarassenko
- Department of Engineering Science, University of Oxford, Oxford OX3 7DQ, U.K
| | - Peter J. Watkinson
- Kadoorie Centre for Critical Care Research and Education, Oxford University Hospitals NHS Foundation Trust, Oxford OX3 9DU, U.K
| | - Richard Beale
- Department of Asthma, Allergy and Lung Biology, King’s College London, London SE1 7EH, U.K
| | - David A. Clifton
- Department of Engineering Science, University of Oxford, Oxford OX3 7DQ, U.K
| |
Collapse
|