Body movement activity recognition for ambulatory cardiac monitoring

Clinical human activity recognition based on a wearable patch of combined tri-axial ACC and ECG sensors

Background

In digital medicine, human activity recognition (HAR) can be used to track and assess a patient's progress throughout rehabilitation, enhancing the quality of life for the elderly and the disabled.

Methods

A patch-type flexible sensor that integrated dynamic electrocardiogram (ECG) and acceleration signal (ACC) was used to record the signals of the various behavioral activities of 20 healthy volunteers and 25 patients with pneumoconiosis. Seven HAR tasks were then carried out on the data using four different deep learning methods (CNN, LSTM, CNN-LSTM and GRU).

Results

When ECG and ACC were obtained simultaneously, the overall accuracy rates of HAR for healthy group were 0.9371, 0.8829, 0.9843 and 0.9486 by the CNN, LSTM, CNN-LSTM and GRU models, respectively. In contrast, the overall accuracy rates of HAR for the pneumoconiosis patients' group were 0.8850, 0.7975, 0.9425 and 0.8525 by the four corresponding models. The accuracy of HAR for both groups using all four models is higher than when only ACC signal is detected.

Conclusion

The addition of the ECG signal significantly improves HAR outcomes in the group of healthy individuals, while having relatively less enhancing effects on the group of patients with pneumoconiosis. When ECG and ACC signals were combined, the increase in HAR accuracy was notable compared to cases where no ECG data was provided. These results suggest that the combination of ACC and ECG data can represent a novel method for the clinical application of HAR.

FaceGuard: A Wearable System To Avoid Face Touching

Most people touch their faces unconsciously, for instance to scratch an itch or to rest one's chin in their hands. To reduce the spread of the novel coronavirus (COVID-19), public health officials recommend against touching one's face, as the virus is transmitted through mucous membranes in the mouth, nose and eyes. Students, office workers, medical personnel and people on trains were found to touch their faces between 9 and 23 times per hour. This paper introduces FaceGuard, a system that utilizes deep learning to predict hand movements that result in touching the face, and provides sensory feedback to stop the user from touching the face. The system utilizes an inertial measurement unit (IMU) to obtain features that characterize hand movement involving face touching. Time-series data can be efficiently classified using 1D-Convolutional Neural Network (CNN) with minimal feature engineering; 1D-CNN filters automatically extract temporal features in IMU data. Thus, a 1D-CNN based prediction model is developed and trained with data from 4,800 trials recorded from 40 participants. Training data are collected for hand movements involving face touching during various everyday activities such as sitting, standing, or walking. Results showed that while the average time needed to touch the face is 1,200 ms, a prediction accuracy of more than 92% is achieved with less than 550 ms of IMU data. As for the sensory response, the paper presents a psychophysical experiment to compare the response time for three sensory feedback modalities, namely visual, auditory, and vibrotactile. Results demonstrate that the response time is significantly smaller for vibrotactile feedback (427.3 ms) compared to visual (561.70 ms) and auditory (520.97 ms). Furthermore, the success rate (to avoid face touching) is also statistically higher for vibrotactile and auditory feedback compared to visual feedback. These results demonstrate the feasibility of predicting a hand movement and providing timely sensory feedback within less than a second in order to avoid face touching.

An Analysis of the Effects of Noisy Electrocardiogram Signal on Heartbeat Detection Performance

Heartbeat detection for ambulatory cardiac monitoring is more challenging as the level of noise and artefacts induced by daily-life activities are considerably higher than monitoring in a hospital setting. It is valuable to understand the relationship between the characteristics of electrocardiogram (ECG) noises and the beat detection performance in the cardiac monitoring system. For this purpose, three well-known algorithms for the beat detection process were re-implemented. The beat detection algorithms were validated using two types of ambulatory datasets, which were the ECG signal from the MIT-BIH Arrhythmia Database and the simulated noise-contaminated ECG signal with different intensities of baseline wander (BW), muscle artefact (MA) and electrode motion (EM) artefact from the MIT-BIH Noise Stress Test Database. The findings showed that signals contaminated with noise and artefacts decreased the potential of beat detection in ambulatory signal with the poorest performance noted for ECG signal affected by the EM artefacts. In conclusion, none of the algorithms was able to detect all QRS complexes without any false detection at the highest level of noise. The EM noise influenced the beat detection performance the most in comparison to the MA and BW noises that resulted in the highest number of misdetections and false detections.

Wearable Sensors for Biochemical Sweat Analysis

Sweat is a largely unexplored biofluid that contains many important biomarkers ranging from electrolytes and metabolites to proteins, cytokines, antigens, and exogenous drugs. The eccrine and apocrine glands produce and excrete sweat through microscale pores on the epidermal surface, offering a noninvasive means for capturing and probing biomarkers that reflect hydration state, fatigue, nutrition, and physiological changes. Recent advances in skin-interfaced wearable sensors capable of real-time in situ sweat collection and analytics provide capabilities for continuous biochemical monitoring in an ambulatory mode of operation. This review presents a broad overview of sweat-based biochemical sensor technologies with an emphasis on enabling materials, designs, and target analytes of interest. The article concludes with a summary of challenges and opportunities for researchers and clinicians in this swiftly growing field.

Classifier for Activities with Variations

Most activity classifiers focus on recognizing application-specific activities that are mostly performed in a scripted manner, where there is very little room for variation within the activity. These classifiers are mainly good at recognizing short scripted activities that are performed in a specific way. In reality, especially when considering daily activities, humans perform complex activities in a variety of ways. In this work, we aim to make activity recognition more practical by proposing a novel approach to recognize complex heterogeneous activities that could be performed in a wide variety of ways. We collect data from 15 subjects performing eight complex activities and test our approach while analyzing it from different aspects. The results show the validity of our approach. They also show how it performs better than the state-of-the-art approaches that tried to recognize the same activities in a more controlled environment.

Application of data fusion techniques and technologies for wearable health monitoring

Technological advances in sensors and communications have enabled discrete integration into everyday objects, both in the home and about the person. Information gathered by monitoring physiological, behavioural, and social aspects of our lives, can be used to achieve a positive impact on quality of life, health, and well-being. Wearable sensors are at the cusp of becoming truly pervasive, and could be woven into the clothes and accessories that we wear such that they become ubiquitous and transparent. To interpret the complex multidimensional information provided by these sensors, data fusion techniques are employed to provide a meaningful representation of the sensor outputs. This paper is intended to provide a short overview of data fusion techniques and algorithms that can be used to interpret wearable sensor data in the context of health monitoring applications. The application of these techniques are then described in the context of healthcare including activity and ambulatory monitoring, gait analysis, fall detection, and biometric monitoring. A snap-shot of current commercially available sensors is also provided, focusing on their sensing capability, and a commentary on the gaps that need to be bridged to bring research to market.

Hierarchical Complex Activity Representation and Recognition Using Topic Model and Classifier Level Fusion

Human activity recognition is an important area of ubiquitous computing. Most current researches in activity recognition mainly focus on simple activities, e.g., sitting, running, walking, and standing. Compared with simple activities, complex activities are more complicated with high-level semantics, e.g., working, commuting, and having a meal. This paper presents a hierarchical model to recognize complex activities as mixtures of simple activities and multiple actions. We generate the components of complex activities using a clustering algorithm, represent and recognize complex activities by applying a topic model on these components. It is a data-driven method that can retain effective information for representing and recognizing complex activities. In addition, acceleration and physiological signals are fused in classifier level to ensure the overall performance of complex activity recognition. The results of experiments show that our method has ability to represent and recognize complex activities effectively.

Comparing wearable devices with wet and textile electrodes for activity recognition

This paper explores the idea of identifying activities from muscle activation which is captured by wearable ECG recording devices that use wet and textile electrodes. Most of the devices available today filter out the high frequency components to retain only the signal related to an ECG. We explain how the high frequency components that correspond to muscle activation can be extracted from the recorded signal and can be used to identify activities. We notice that is possible to obtain good performance for both the wet and dry electrodes. However, we observed that signals from the dry textile electrodes introduce less artifacts associated with muscle activation.

Low energy physical activity recognition system on smartphones

An innovative approach to physical activity recognition based on the use of discrete variables obtained from accelerometer sensors is presented. The system first performs a discretization process for each variable, which allows efficient recognition of activities performed by users using as little energy as possible. To this end, an innovative discretization and classification technique is presented based on the χ² distribution. Furthermore, the entire recognition process is executed on the smartphone, which determines not only the activity performed, but also the frequency at which it is carried out. These techniques and the new classification system presented reduce energy consumption caused by the activity monitoring system. The energy saved increases smartphone usage time to more than 27 h without recharging while maintaining accuracy.

False alarm reduction in BSN-based cardiac monitoring using signal quality and activity type information

False alarms in cardiac monitoring affect the quality of medical care, impacting on both patients and healthcare providers. In continuous cardiac monitoring using wireless Body Sensor Networks (BSNs), the quality of ECG signals can be deteriorated owing to several factors, e.g., noises, low battery power, and network transmission problems, often resulting in high false alarm rates. In addition, body movements occurring from activities of daily living (ADLs) can also create false alarms. This paper presents a two-phase framework for false arrhythmia alarm reduction in continuous cardiac monitoring, using signals from an ECG sensor and a 3D accelerometer. In the first phase, classification models constructed using machine learning algorithms are used for labeling input signals. ECG signals are labeled with heartbeat types and signal quality levels, while 3D acceleration signals are labeled with ADL types. In the second phase, a rule-based expert system is used for combining classification results in order to determine whether arrhythmia alarms should be accepted or suppressed. The proposed framework was validated on datasets acquired using BSNs and the MIT-BIH arrhythmia database. For the BSN dataset, acceleration and ECG signals were collected from 10 young and 10 elderly subjects while they were performing ADLs. The framework reduced the false alarm rate from 9.58% to 1.43% in our experimental study, showing that it can potentially assist physicians in diagnosing a vast amount of data acquired from wireless sensors and enhance the performance of continuous cardiac monitoring.

Study on a real-time BEAM system for diagnosis assistance based on a system on chips design

As an innovative as well as an interdisciplinary research project, this study performed an analysis of brain signals so as to establish BrainIC as an auxiliary tool for physician diagnosis. Cognition behavior sciences, embedded technology, system on chips (SOC) design and physiological signal processing are integrated in this work. Moreover, a chip is built for real-time electroencephalography (EEG) processing purposes and a Brain Electrical Activity Mapping (BEAM) system, and a knowledge database is constructed to diagnose psychosis and body challenges in learning various behaviors and signals antithesis by a fuzzy inference engine. This work is completed with a medical support system developed for the mentally disabled or the elderly abled.

A novel application for the detection of an irregular pulse using an iPhone 4S in patients with atrial fibrillation

BACKGROUND

Atrial fibrillation (AF) is common and associated with adverse health outcomes. Timely detection of AF can be challenging using traditional diagnostic tools. Smartphone use is increasing and may provide an inexpensive and user-friendly means to diagnoseAF.

OBJECTIVE

To test the hypothesis that a smartphone-based application could detect an irregular pulse fromAF.

METHODS

Seventy-six adults with persistent AF were consented for participation in our study. We obtained pulsatile time series recordings before and after cardioversion using an iPhone 4S camera. A novel smartphone application conducted real-time pulse analysis using 2 statistical methods: root mean square of successive RR difference (RMSSD/mean) and Shannon entropy (ShE). We examined the sensitivity, specificity, and predictive accuracy of both algorithms using the 12-lead electrocardiogram as the gold standard.

RESULTS

RMSDD/mean and ShE were higher in participants in AF than in those with sinus rhythm. The 2 methods were inversely related to AF in regression models adjusting for key factors including heart rate and blood pressure (beta coefficients per SD increment in RMSDD/mean and ShE were-0.20 and-0.35; P<.001). An algorithm combining the 2 statistical methods demonstrated excellent sensitivity (0.962), specificity (0.975), and accuracy (0.968) for beat-to-beat discrimination of an irregular pulse during AF from sinus rhythm.

CONCLUSIONS

In a prospectively recruited cohort of 76 participants undergoing cardioversion for AF, we found that a novel algorithm analyzing signals recorded using an iPhone 4S accurately distinguished pulse recordings during AF from sinus rhythm. Data are needed to explore the performance and acceptability of smartphone-based applications for AF detection.

RPCA-based detection and quantification of motion artifacts in ECG signals

In this paper, a recursive principal component analysis (RPCA)-based algorithm is applied for detecting and quantifying the motion artifact episodes encountered in an ECG signal. The motion artifact signal is synthesized by low-pass filtering a random noise signal with different spectral ranges of LPF (low pass filter): 0-5 Hz, 0-10 Hz, 0-15 Hz and 0-20 Hz. Further, the analysis of the algorithm is carried out for different values of SNR levels and forgetting factors (α) of an RPCA algorithm. The algorithm derives an error signal, wherever a motion artifact episode (noise) is present in the entire ECG signal with 100% accuracy. The RPCA error magnitude is almost zero for the clean signal portion and considerably high wherever the motion artifacts (noisy episodes) are encountered in the ECG signals. Further, the general trend of the algorithm is to produce a smaller magnitude of error for higher SNR (i.e. low level of noise) and vice versa. The quantification of the RPCA algorithm has been made by applying it over 25 ECG data-sets of different morphologies and genres with three different values of SNRs for each forgetting factor and for each of four spectral ranges.

EVALUATION OF QRS COMPLEX BASED ON DWT COEFFICIENTS ANALYSIS USING DAUBECHIES WAVELETS FOR DETECTION OF MYOCARDIAL ISCHAEMIA

This paper presents a new approach in the field of electrocardiogram (ECG) feature extraction system based on the discrete wavelet transform (DWT) coefficients using Daubechies Wavelets. Real ECG signals recorded in lead II configuration are chosen for processing. The ECG signal was acquired by a battery operated, portable ECG data acquisition and signal processing module. In the second step the ECG signal was denoised using soft thresholding with Symlet4 wavelet. Further denoising was achieved by removing the corresponding wavelet coefficients at higher levels of decomposition. Later the ECG data files were converted to .txt files and subsequently to. mat files before being imported into the Matlab 7.4.0 environment for the computation of the decomposition coefficients. The QRS complexes were grouped as normal or myocardial ischaemic ones based on these decomposition coefficients. The algorithm developed by us was evaluated with control database comprising 120 records and validated using 60 records making up test database. By using the DWT coefficients, we have successfully achieved the myocardial ischaemia detection rates up to 97.5% with the technique developed by us for control data and up to 100% for validation test data.

EMERGENCY HEALTH CARE AND FOLLOW-UP TELEMEDICINE SYSTEM FOR RURAL AREAS BASED ON LABVIEW

This paper presents the design and development of a prototype for remote ECG data transmission based on Internet-enabled health care services and telemedicine fundamentals. An ECG acquisition system developed by the authors is used to acquire the ECG signal in lead-II configuration from patient and store it in .lvm format in a PC interfaced to patient module through RS232. This unit (data server) on the patient side then transfers the data to a remote client (on doctor's side) using TCP/IP as network protocol on LabVIEW 8.20 environment. Using this device, a specialist doctor can telematically move to the patient site and instruct medical personnel when handling a patient. During the last years, more and more modern tools have found their ways to different tasks during the design, the realization, and data processing in the area of Internet access to the e-health services. The telemedicine system demonstrated in this work is a combined real-time and store-and-forward facility.

Collaborative processing of wearable and ambient sensor system for blood pressure monitoring

This paper describes wireless wearable and ambient sensors that cooperate to monitor a person’s vital signs such as heart rate and blood pressure during daily activities. Each wearable sensor is attached on different parts of the body. The wearable sensors require a high sampling rate and time synchronization to provide a precise analysis of the received signals. The trigger signal for synchronization is provided by the ambient sensors, which detect the user’s presence. The Bluetooth and IEEE 802.15.4 wireless technologies are used for real-time sensing and time synchronization. Thus, this wearable health-monitoring sensor response is closely related to the context in which it is being used. Experimental results indicate that the system simultaneously provides information about the user’s location and vital signs, and the synchronized wearable sensors successfully measures vital signs with a 1 ms resolution.

Multimodal physical activity recognition by fusing temporal and cepstral information

A physical activity (PA) recognition algorithm for a wearable wireless sensor network using both ambulatory electrocardiogram (ECG) and accelerometer signals is proposed. First, in the time domain, the cardiac activity mean and the motion artifact noise of the ECG signal are modeled by a Hermite polynomial expansion and principal component analysis, respectively. A set of time domain accelerometer features is also extracted. A support vector machine (SVM) is employed for supervised classification using these time domain features. Second, motivated by their potential for handling convolutional noise, cepstral features extracted from ECG and accelerometer signals based on a frame level analysis are modeled using Gaussian mixture models (GMMs). Third, to reduce the dimension of the tri-axial accelerometer cepstral features which are concatenated and fused at the feature level, heteroscedastic linear discriminant analysis is performed. Finally, to improve the overall recognition performance, fusion of the multimodal (ECG and accelerometer) and multidomain (time domain SVM and cepstral domain GMM) subsystems at the score level is performed. The classification accuracy ranges from 79.3% to 97.3% for various testing scenarios and outperforms the state-of-the-art single accelerometer based PA recognition system by over 24% relative error reduction on our nine-category PA database.

A survey on wearable biosensor systems for health monitoring

Wearable biosensor systems for health monitoring are an emerging trend and are expected to enable proactive personal health management and better treatment of various medical conditions. These systems, comprising various types of small physiological sensors, transmission modules and processing capabilities, promise to change the future of health care, by providing low-cost wearable unobtrusive solutions for continuous all-day and any-place health, mental and activity status monitoring. This paper presents a comprehensive survey on the research and development done so far on wearable biosensor systems for health-monitoring, by comparing a variety of current system implementations and approaches and identifying their technological shortcomings. A set of significant features, that best describe the functionality and the characteristics of wearable biosensor systems, has been selected to derive a thorough study. The aim of this survey is not to criticize, but to serve as a reference for current achievements and their maturity level and to provide direction for future research improvements.

Impact analysis of body movement in ambulatory ECG

Ambulatory ECG analysis is adversely affected by motion artifacts induced due to body movements. Knowledge of the extent of motion artifacts facilitates better ECG analysis. In [1], an unsupervised method using recursive principal component analysis (RPCA) was used to detect transitions between body movements. In this paper, we endeavour to quantify the impact of various types of body movements on the extent of ECG motion artifact using the RPCA error signal. For this purpose, acceleration data from different body parts i.e. arm(s), leg and waist, have been obtained using commercially available motion sensors, in conjunction with ECG signal, while carrying out routine body movement activities like climbing stairs, walking, twisting, and arm movements, at three different intensity levels: slow, medium and fast. The acceleration magnitudes and the RPCA error sequence are found to be well correlated, thus validating the body movement impact analysis, and also indicating the suitability of the method for quantification of body movement kinematics from the ECG signal itself in the absence of any accelerometer sensors.

Transition Detection in Body Movement Activities for Wearable ECG

It has been shown by Pawar et al. (2007) that the motion artifacts induced by body movement activity (BMA) in a single-lead wearable electrocardiogram (ECG) signal recorder, while monitoring an ambulatory patient, can be detected and removed by using a principal component analysis (PCA)-based classification technique. However, this requires the ECG signal to be temporally segmented so that each segment comprises of artifacts due to a single type of body movement activity. In this paper, we propose a simple, recursively updated PCA-based technique to detect transitions wherever the type of body movement is changed.