Quantitative Analysis of Mother Wavelet Function Selection for Wearable Sensors-Based Human Activity Recognition.

Recent advancements in the Internet of Things (IoT) wearable devices such as wearable inertial sensors have increased the demand for precise human activity recognition (HAR) with minimal computational resources. The wavelet transform, which offers excellent time-frequency localization characteristics, is well suited for HAR recognition systems. Selecting a mother wavelet function in wavelet analysis is critical, as optimal selection improves the recognition performance. The activity time signals data have different periodic patterns that can discriminate activities from each other. Therefore, selecting a mother wavelet function that closely resembles the shape of the recognized activity's sensor (inertial) signals significantly impacts recognition performance. This study uses an optimal mother wavelet selection method that combines wavelet packet transform with the energy-to-Shannon-entropy ratio and two classification algorithms: decision tree (DT) and support vector machines (SVM). We examined six different mother wavelet families with different numbers of vanishing points. Our experiments were performed on eight publicly available ADL datasets: MHEALTH, WISDM Activity Prediction, HARTH, HARsense, DaLiAc, PAMAP2, REALDISP, and HAR70+. The analysis demonstrated in this paper can be used as a guideline for optimal mother wavelet selection for human activity recognition.

Signal Quality Assessment of Compressively Sensed Electrocardiogram.

OBJECTIVE: Develop a signal quality index (SQI) to determine the quality of compressively sensed electrocardiogram (ECG) by estimating the signal-to-noise ratio (SNR). METHODS: The SQI used random forests, with the ratio of the standard deviations of an ECG segment and a clean ECG and the Wasserstein metric between the amplitude distributions of an ECG segment and a clean ECG, as features. The SQI was tested using the Long-Term Atrial Fibrillation Database (LTAFDB) and the PhysioNet/CinC Challenge 2011 Database Set A (CinCDB). Clean ECG segments from the LTAFDB were corrupted using simulated motion artifact, with preset SNR between -12 dB and 12 dB. The CinCDB was used as-it-is. The databases were compressively sensed using three types of sensing matrices at three compression ratios (50%, 75%, and 95%). For LTAFDB, the RMSE and Spearman correlation between the SQI and the preset SNR were used for evaluation, while for CinCDB, accuracy and F1 score were used. RESULTS: The average RMSE was 3.18 dB and 3.47 dB in normal and abnormal ECG. The average Spearman correlation was 0.94 and 0.93 in normal and abnormal ECG, respectively. The average accuracy and F1 score were 0.90 and 0.88, respectively. CONCLUSION: The SQI determined the quality of compressively sensed ECG and generalized across different databases. There was no consequential effect on the SQI due to abnormal ECG or compression using different sensing matrices and compression ratios. SIGNIFICANCE: Without reconstruction, the SQI can inform which ECG should be analyzed to reduce false alarms due to contamination.

Image Super-Resolution Through Compressive Sensing-based Recovery.

The primary aim of image super-resolution techniques is to produce a high resolution (HR) image from a low resolution (LR) image efficiently. Deep learning algorithms are being extensively used to address the ill-posed problem of single image super-resolution which requires extremely large data-sets and high processing power. When one does not have access to large data-sets or have limited processing power, an alternative technique may be in order. In this study, we have developed a novel positive scale image resizing method inspired by compressive sensing (CS). We have considered the image super-resolution as a CS recovery problem in which a low resolution image is assumed as a compressed measurement and the required interpolated image is treated as output of the CS-based recovery. In the proposed HR recovery method, a deterministic binary block diagonal measurement matrix, (DBBD), is used as measurement matrix since it maintains the visual similarity between the low and high resolution images. Then along with a sparsification matrix, the sparse representation of HR image is first recovered and subsequently the dense HR image is obtained. The proposed method is applied to medical and non-medical images. The HR images obtained using the traditional proximal, bilinear and bi-cubic interpolation techniques are compared with those obtained using the proposed method. The proposed CS inspired method delivers superior HR images than the traditional techniques. The superiority of the proposed method is attributed to the unique usage of the DBBD matrix and the CS recovery algorithm to obtain a high resolution image without any prior training data-set.

An Investigation of Rotary Drone HERM Line Spectrum under Manoeuvering Conditions.

Detecting and identifying drones is of great interest due to the proliferation of highly manoeuverable drones with on-board sensors of increasing sensing capabilities. In this paper, we investigate the use of radars for tackling this problem. In particular, we focus on the problem of detecting rotary drones and distinguishing between single-propeller and multi-propeller drones using a micro-Doppler analysis. Two different radars were used, an ultra wideband (UWB) continuous wave (CW) C-band radar and an automotive frequency modulated continuous wave (FMCW) W-band radar, to collect micro-Doppler signatures of the drones. By taking a closer look at HElicopter Rotor Modulation (HERM) lines, the spool and chopping lines are identified for the first time in the context of drones to determine the number of propeller blades. Furthermore, a new multi-frequency analysis method using HERM lines is developed, which allows the detection of propeller rotation rates (spool and chopping frequencies) of single and multi-propeller drones. Therefore, the presented method is a promising technique to aid in the classification of drones.

Transfer Learning for Detection of Atrial Fibrillation in Deterministic Compressive Sensed ECG.

Atrial Fibrillation (AF) is a cardiac condition resulting from uncoordinated contraction of the atria which may lead to an increase in the risk of heart attacks, strokes, and death. AF symptoms may go undetected and may require longterm monitoring of electrocardiogram (ECG) to be detected. Long-term ECG monitoring can generate a large amount of data which can increase power, storage, and the wireless transmission bandwidth of monitoring devices. Compressive Sensing (CS) is compression technique at the sampling stage which may save power, storage, and wireless bandwidth of monitoring devices. The reconstruction of compressive sensed ECG is a computationally expensive operation; therefore, detection of AF in compressive sensed ECG is warranted. This paper presents preliminary results of using deep learning to detect AF in deterministic compressive sensed ECG. MobileNetV2 convolutional neural network (CNN) was used in this paper. Transfer learning was utilized to leverage a pre-trained CNN with the final two layers retrained using 24 records from the Long-Term Atrial Fibrillation Database. The Short-Term Fourier Transform was used to generate spectrograms that were fed to the CNN. The CNN was tested on the MIT-BIH Atrial Fibrillation Database at the uncompressed, 50%, 75%, and 95% compressed ECG. The performance of the CNN was evaluated using weighted average precision (AP) and area under the curve (AUC) of the receiver operator curve (ROC). The CNN had AP of 0.80, 0.70, 0.70, and 0.57 at uncompressed, 50%, 75%, and 95% compression levels. The AUC was 0.87, 0.78, 0.79, and 0.75 at each compression level. The preliminary results show promise for using deep learning to detect AF in compressive sensed ECG.Clinical Relevance-This paper confirms that AF can be detected in compressive sensed ECG using deep learning, This will facilitate long-term ECG monitoring using wearable devices and will reduce adverse complications resulting from undiagnosed AF.

Uncertainty in Blood Pressure Measurement Estimated Using Ensemble-Based Recursive Methodology.

Automated oscillometric blood pressure monitors are commonly used to measure blood pressure for many patients at home, office, and medical centers, and they have been actively studied recently. These devices usually provide a single blood pressure point and they are not able to indicate the uncertainty of the measured quantity. We propose a new technique using an ensemble-based recursive methodology to measure uncertainty for oscillometric blood pressure measurements. There are three stages we consider: the first stage is pre-learning to initialize good parameters using the bagging technique. In the second stage, we fine-tune the parameters using the ensemble-based recursive methodology that is used to accurately estimate blood pressure and then measure the uncertainty for the systolic blood pressure and diastolic blood pressure in the third stage.

Evaluation of the signal quality of wrist-based photoplethysmography.

OBJECTIVE: Wearable devices with embedded photoplethysmography (PPG) sensors enable continuous monitoring of cardiovascular activity, allowing for the detection cardiovascular problems, such as arrhythmias. However, the quality of wrist-based PPG is highly variable, and is subject to artifacts from motion and other interferences. The goal of this paper is to evaluate the signal quality obtained from wrist-based PPG when used in an ambulatory setting. APPROACH: Ambulatory data were collected over a 24 h period for 10 elderly, and 16 non-elderly participants. Visual assessment is used as the gold standard for PPG signal quality, with inter-rater agreement evaluated using Fleiss' Kappa. With this gold standard, 5 classifiers were evaluated using a modified 13-fold cross-validation approach. MAIN RESULTS: A Random Forest quality classification algorithm showed the best performance, with an accuracy of 74.5%, and was then used to evaluate 24 h long ambulatory wrist-based PPG measurements. SIGNIFICANCE: In general, data quality was high at night, and low during the day. Our results suggest wrist-based PPG may be best for continuous cardiovascular monitoring applications during the night, but less useful during the day unless methods can be identified to improve low quality signal segments.

Classification of Human Posture from Radar Returns Using Ultra-Wideband Radar.

There is a great need for new technology that helps ensure the well-being of senior citizens who have compromised health and are at an elevated risk of injury due to falls. Being able to detect posture and postural changes may be helpful in prediction and prevention of impending falls. Ultra-Wideband (UWB) radar is an attractive means for patient monitoring because it is inexpensive, capable of penetrating obstacles, privacy preserving and it consumes little power. In this paper, classification of postures, namely sitting, standing and lying is presented using stand-off sensing using UWB radar in an indoor environment. It is found that using location specific classifiers, overall accuracy can be improved. In this paper, a decision tree classifier capable of achieving 85% overall accuracy is proposed. This classifier uses 33 features from 10 second data sample segments.

Event Recognition for Contactless Activity Monitoring Using Phase-Modulated Continuous Wave Radar.

OBJECTIVES: The use of remote sensing technologies such as radar is gaining popularity as a technique for contactless detection of physiological signals and analysis of human motion. This paper presents a methodology for classifying different events in a collection of phase modulated continuous wave radar returns. The primary application of interest is to monitor inmates where the presence of human vital signs amidst different, interferences needs to be identified. METHODS: A comprehensive set of features is derived through time and frequency domain analyses of the radar returns. The Bhattacharyya distance is used to preselect the features with highest class separability as the possible candidate features for use in the classification process. The uncorrelated linear discriminant analysis is performed to decorrelate, denoise, and reduce the dimension of the candidate feature set. Linear and quadratic Bayesian classifiers are designed to distinguish breathing, different human motions, and nonhuman motions. The performance of these classifiers is evaluated on a pilot dataset of radar returns that contained different events including breathing, stopped breathing, simple human motions, and movement of fan and water. RESULTS: Our proposed pattern classification system achieved accuracies of up to 93% in stationary subject detection, 90% in stop-breathing detection, and 86% in interference detection. CONCLUSION: Our proposed radar pattern recognition system was able to accurately distinguish the predefined events amidst interferences. SIGNIFICANCE: Besides inmate monitoring and suicide attempt detection, this paper can be extended to other radar applications such as home-based monitoring of elderly people, apnea detection, and home occupancy detection.

Oscillometric blood pressure estimation by combining nonparametric bootstrap with Gaussian mixture model.

BACKGROUND: Blood pressure (BP) is one of the most important vital indicators and plays a key role in determining the cardiovascular activity of patients. METHODS: This paper proposes a hybrid approach consisting of nonparametric bootstrap (NPB) and machine learning techniques to obtain the characteristic ratios (CR) used in the blood pressure estimation algorithm to improve the accuracy of systolic blood pressure (SBP) and diastolic blood pressure (DBP) estimates and obtain confidence intervals (CI). The NPB technique is used to circumvent the requirement for large sample set for obtaining the CI. A mixture of Gaussian densities is assumed for the CRs and Gaussian mixture model (GMM) is chosen to estimate the SBP and DBP ratios. The K-means clustering technique is used to obtain the mixture order of the Gaussian densities. RESULTS: The proposed approach achieves grade "A" under British Society of Hypertension testing protocol and is superior to the conventional approach based on maximum amplitude algorithm (MAA) that uses fixed CR ratios. The proposed approach also yields a lower mean error (ME) and the standard deviation of the error (SDE) in the estimates when compared to the conventional MAA method. In addition, CIs obtained through the proposed hybrid approach are also narrower with a lower SDE. CONCLUSIONS: The proposed approach combining the NPB technique with the GMM provides a methodology to derive individualized characteristic ratio. The results exhibit that the proposed approach enhances the accuracy of SBP and DBP estimation and provides narrower confidence intervals for the estimates.

Bayesian fusion algorithm for improved oscillometric blood pressure estimation.

A variety of oscillometric algorithms have been recently proposed in the literature for estimation of blood pressure (BP). However, these algorithms possess specific strengths and weaknesses that should be taken into account before selecting the most appropriate one. In this paper, we propose a fusion method to exploit the advantages of the oscillometric algorithms and circumvent their limitations. The proposed fusion method is based on the computation of the weighted arithmetic mean of the oscillometric algorithms estimates, and the weights are obtained using a Bayesian approach by minimizing the mean square error. The proposed approach is used to fuse four different oscillometric blood pressure estimation algorithms. The performance of the proposed method is evaluated on a pilot dataset of 150 oscillometric recordings from 10 subjects. It is found that the mean error and standard deviation of error are reduced relative to the individual estimation algorithms by up to 7 mmHg and 3 mmHg in estimation of systolic pressure, respectively, and by up to 2 mmHg and 3 mmHg in estimation of diastolic pressure, respectively.

Oscillometric Blood Pressure Estimation: Past, Present, and Future.

The use of automated blood pressure (BP) monitoring is growing as it does not require much expertise and can be performed by patients several times a day at home. Oscillometry is one of the most common measurement methods used in automated BP monitors. A review of the literature shows that a large variety of oscillometric algorithms have been developed for accurate estimation of BP but these algorithms are scattered in many different publications or patents. Moreover, considering that oscillometric devices dominate the home BP monitoring market, little effort has been made to survey the underlying algorithms that are used to estimate BP. In this review, a comprehensive survey of the existing oscillometric BP estimation algorithms is presented. The survey covers a broad spectrum of algorithms including the conventional maximum amplitude and derivative oscillometry as well as the recently proposed learning algorithms, model-based algorithms, and algorithms that are based on analysis of pulse morphology and pulse transit time. The aim is to classify the diverse underlying algorithms, describe each algorithm briefly, and discuss their advantages and disadvantages. This paper will also review the artifact removal techniques in oscillometry and the current standards for the automated BP monitors.

Estimated confidence interval from single blood pressure measurement based on algorithmic fusion.

BACKGROUND: Current oscillometric blood pressure measurement devices generally provide only single-point estimates for systolic and diastolic blood pressures and rarely provide confidence ranges for these estimates. A novel methodology to obtain confidence intervals (CIs) for systolic blood pressure (SBP) and diastolic blood pressure (DBP) estimates from a single oscillometric blood pressure measurement is presented. METHODS: The proposed methodology utilizes the multiple regression technique to fuse optimally a set of SBP and DBP estimates obtained through different algorithms. However, the set of SBP and DBP estimates is a small number to determine the CI of each individual subject. To address this issue, the weighted bootstrap approach based on the multiple regression technique was used to generate a pseudo sample set for the SBP and the DBP. In this paper, the multiple regression technique can estimate the best-fitting surface of an efficient function that relates the input sample set as an independent vector to the auscultatory nurse measurement as a dependent vector to estimate regression coefficients. Consequently, the coefficients are assigned to an eight-sample set obtained from the fusion of different algorithms as optimally weighted parameters. CIs are also estimated using the conventional methods on the set of fused SBP and DBP estimates for comparison purposes. RESULTS: The proposed method was applied to an experimental dataset of 85 patients. The results indicated that the proposed approach provides better blood pressure estimates than the existing algorithms and, in addition, is able to provide CIs for a single measurement. CONCLUSIONS: The CIs derived from the proposed scheme are much smaller than those calculated by conventional methods except for the pseudo maximum amplitude-envelope algorithm for both the SBP and the DBP, probably because of the decrease in the standard deviation through the increase in the pseudo measurements using the weighted bootstrap method for each subject. The proposed methodology is likely the only one currently available that can provide CIs for single-sample blood pressure measurements.

Model of human breathing reflected signal received by PN-UWB radar.

Human detection is an integral component of civilian and military rescue operations, military surveillance and combat operations. Human detection can be achieved through monitoring of vital signs. In this article, a mathematical model of human breathing reflected signal received in PN-UWB radar is proposed. Unlike earlier published works, both chest and abdomen movements are considered for modeling the radar return signal along with the contributions of fundamental breathing frequency and its harmonics. Analyses of recorded reflected signals from three subjects in different postures and at different ranges from the radar indicate that ratios of the amplitudes of the harmonics contain information about posture and posture change.

Blood pressure estimation using maximum slope of oscillometric pulses.

A new oscillometric pulse index (OPI) derived from the maximum slope (MS) of each pulse in the oscillometric blood pressure waveform is proposed for blood pressure estimation. Maximum slope for each pulse is obtained using the first derivative of the pulse and an envelope of the values corresponding to the maximum slopes is obtained. The maximum of the envelope is taken as the mean arterial pressure (MAP) and the systolic blood pressure (SBP) and diastolic blood pressure (DBP) estimates are obtained as a fraction of the MAP, similar to the traditional maximum amplitude algorithm (MAA). The proposed algorithm is tested on 18 healthy subjects. The MAP, SBP and DBP estimates obtained from the proposed algorithm are compared with those obtained from a commercial blood pressure device and with the estimates obtained using the MAA and morphological qualitative measures available in the literature.

Blood pressure estimation using oscillometric pulse morphology.

A novel pulse morphology-based approach for estimation of blood pressure from non-invasive oscillometric blood pressure measurement is presented. Quantitative measures that describe the pulse morphology are utilized to obtain the estimates of mean arterial, systolic, and diastolic pressures. Preliminary results obtained from a small set of measurements are used to demonstrate the feasibility of the proposed approach. The estimates obtained through pulse morphology-based approach is compared with those obtained from a commercial blood pressure device.