Research

Jump to: Other

22 pages, 6997 KiB

Open AccessArticle

Combining Supervised and Unsupervised Learning Algorithms for Human Activity Recognition

by Elena-Alexandra Budisteanu and Irina Georgiana Mocanu

Sensors 2021, 21(18), 6309; https://0-doi-org.brum.beds.ac.uk/10.3390/s21186309 - 21 Sep 2021

Cited by 4 | Viewed by 2658

Human activity recognition is an extensively researched topic in the last decade. Recent methods employ supervised and unsupervised deep learning techniques in which spatial and temporal dependency is modeled. This paper proposes a novel approach for human activity recognition using skeleton data. The [...] Read more.

Human activity recognition is an extensively researched topic in the last decade. Recent methods employ supervised and unsupervised deep learning techniques in which spatial and temporal dependency is modeled. This paper proposes a novel approach for human activity recognition using skeleton data. The method combines supervised and unsupervised learning algorithms in order to provide qualitative results and performance in real time. The proposed method involves a two-stage framework: the first stage applies an unsupervised clustering technique to group up activities based on their similarity, while the second stage classifies data assigned to each group using graph convolutional networks. Different clustering techniques and data augmentation strategies are explored for improving the training process. The results were compared against the state of the art methods and the proposed model achieved 90.22% Top-1 accuracy performance for NTU-RGB+D dataset (the performance was increased by approximately 9% compared with the baseline graph convolutional method). Moreover, inference time and total number of parameters stay within the same magnitude order. Extending the initial set of activities with additional classes is fast and robust, since there is no required retraining of the entire architecture but only to retrain the cluster to which the activity is assigned. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

29 pages, 12119 KiB

Open AccessArticle

A Novel Multi-Sensor Fusion Algorithm Based on Uncertainty Analysis

by Haobai Xue, Maomao Zhang, Peining Yu, Haifeng Zhang, Guozhu Wu, Yi Li and Xiangyuan Zheng

Sensors 2021, 21(8), 2713; https://0-doi-org.brum.beds.ac.uk/10.3390/s21082713 - 12 Apr 2021

Cited by 7 | Viewed by 2281

Abstract

During the research and development of multiphase flowmeters, errors are often used to evaluate the advantages and disadvantages of different devices and algorithms, whilst an in-depth uncertainty analysis is seldom carried out. However, limited information is sometimes revealed from the errors, especially when [...] Read more.

During the research and development of multiphase flowmeters, errors are often used to evaluate the advantages and disadvantages of different devices and algorithms, whilst an in-depth uncertainty analysis is seldom carried out. However, limited information is sometimes revealed from the errors, especially when the test data are scant, and this makes an in-depth comparison of different algorithms impossible. In response to this problem, three combinations of sensing methods are implemented, which are the “capacitance and cross-correlation”, the “cross-correlation and differential pressure” and the “differential pressure and capacitance” respectively. The analytical expressions of the gas/liquid flowrate and the associated standard uncertainty have been derived, and Monte Carlo simulations are carried out to determine the desired probability density function. The results obtained through these two approaches are basically the same. Thereafter, the sources of uncertainty for each combination are traced and their respective variations with flowrates are analyzed. Further, the relationship between errors and uncertainty is studied, which demonstrates that the two uncertainty analysis approaches can be a powerful tool for error prediction. Finally, a novel multi-sensor fusion algorithm based on the uncertainty analysis is proposed. This algorithm can minimize the standard uncertainty over the whole flowrate range and thus reduces the measurement error. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

10 pages, 3290 KiB

Open AccessCommunication

Blood Pressure Monitoring System Using a Two-Channel Ballistocardiogram and Convolutional Neural Networks

by Woojoon Seok, Kwang Jin Lee, Dongrae Cho, Jongryun Roh and Sayup Kim

Sensors 2021, 21(7), 2303; https://0-doi-org.brum.beds.ac.uk/10.3390/s21072303 - 25 Mar 2021

Cited by 18 | Viewed by 2943

Abstract

Hypertension is a chronic disease that kills 7.6 million people worldwide annually. A continuous blood pressure monitoring system is required to accurately diagnose hypertension. Here, a chair-shaped ballistocardiogram (BCG)-based blood pressure estimation system was developed with no sensors attached to users. Two experimental [...] Read more.

Hypertension is a chronic disease that kills 7.6 million people worldwide annually. A continuous blood pressure monitoring system is required to accurately diagnose hypertension. Here, a chair-shaped ballistocardiogram (BCG)-based blood pressure estimation system was developed with no sensors attached to users. Two experimental sessions were conducted with 30 subjects. In the first session, two-channel BCG and blood pressure data were recorded for each subject. In the second session, the two-channel BCG and blood pressure data were recorded after running on a treadmill and then resting on the newly developed system. The empirical mode decomposition algorithm was used to remove noise in the two-channel BCG, and the instantaneous phase was calculated by applying a Hilbert transform to the first intrinsic mode functions. After training a convolutional neural network regression model that predicts the systolic and diastolic blood pressures (SBP and DBP) from the two-channel BCG phase, the results of the first session (rest) and second session (recovery) were compared. The results confirmed that the proposed model accurately estimates the rapidly rising blood pressure in the recovery state. Results from the rest sessions satisfied the Association for the Advancement of Medical Instrumentation (AAMI) international standards. The standard deviation of the SBP results in the recovery session exceeded 0.7. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

31 pages, 1840 KiB

Open AccessArticle

Hepatocellular Carcinoma Automatic Diagnosis within CEUS and B-Mode Ultrasound Images Using Advanced Machine Learning Methods

by Delia Mitrea, Radu Badea, Paulina Mitrea, Stelian Brad and Sergiu Nedevschi

Sensors 2021, 21(6), 2202; https://0-doi-org.brum.beds.ac.uk/10.3390/s21062202 - 21 Mar 2021

Cited by 15 | Viewed by 6078

Abstract

Hepatocellular Carcinoma (HCC) is the most common malignant liver tumor, being present in 70% of liver cancer cases. It usually evolves on the top of the cirrhotic parenchyma. The most reliable method for HCC diagnosis is the needle biopsy, which is an invasive, [...] Read more.

Hepatocellular Carcinoma (HCC) is the most common malignant liver tumor, being present in 70% of liver cancer cases. It usually evolves on the top of the cirrhotic parenchyma. The most reliable method for HCC diagnosis is the needle biopsy, which is an invasive, dangerous method. In our research, specific techniques for non-invasive, computerized HCC diagnosis are developed, by exploiting the information from ultrasound images. In this work, the possibility of performing the automatic diagnosis of HCC within B-mode ultrasound and Contrast-Enhanced Ultrasound (CEUS) images, using advanced machine learning methods based on Convolutional Neural Networks (CNN), was assessed. The recognition performance was evaluated separately on B-mode ultrasound images and on CEUS images, respectively, as well as on combined B-mode ultrasound and CEUS images. For this purpose, we considered the possibility of combining the input images directly, performing feature level fusion, then providing the resulted data at the entrances of representative CNN classifiers. In addition, several multimodal combined classifiers were experimented, resulted by the fusion, at classifier, respectively, at the decision levels of two different branches based on the same CNN architecture, as well as on different CNN architectures. Various combination methods, and also the dimensionality reduction method of Kernel Principal Component Analysis (KPCA), were involved in this process. These results were compared with those obtained on the same dataset, when employing advanced texture analysis techniques in conjunction with conventional classification methods and also with equivalent state-of-the-art approaches. An accuracy above 97% was achieved when our new methodology was applied. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

19 pages, 4173 KiB

Open AccessArticle

A Wet Gas Metering System Based on the Extended-Throat Venturi Tube

by Haobai Xue, Peining Yu, Maomao Zhang, Haifeng Zhang, Encheng Wang, Guozhu Wu, Yi Li and Xiangyuan Zheng

Sensors 2021, 21(6), 2120; https://0-doi-org.brum.beds.ac.uk/10.3390/s21062120 - 17 Mar 2021

Cited by 7 | Viewed by 2072

Abstract

Although the use of a classical Venturi tube for wet gas metering has been extensively studied in the literature, the use of an extended-throat Venturi (ETV) tube has rarely been reported since its first proposal by J. R. Fincke in 1999. The structure [...] Read more.

Although the use of a classical Venturi tube for wet gas metering has been extensively studied in the literature, the use of an extended-throat Venturi (ETV) tube has rarely been reported since its first proposal by J. R. Fincke in 1999. The structure of an ETV is very simple, but due to the complexity of multiphase flow, its theoretical model has not been fully established yet. Therefore, in this paper theoretical models have been developed for the convergent and throat sections of an ETV, and the gradients of front and rear differential pressures are derived analytically. Several flowrate algorithms have been proposed and compared with the existing ones. Among them, the iteration algorithm is found to be the best. A reasonable explanation is provided for its performance. The relationship between the differential pressure gradient and the flowrate relative error is also studied, such that the relative error distributions varying with ETV measured flowrates can be derived. The gas flowrate error of ETV increases with the liquid content whilst the liquid flowrate error of ETV decreases with the liquid content, and the relative errors of liquid flowrate are generally 2 to 3 times larger than that of the gas flowrate. Finally, the ETV tends to be more accurate than the classical Venturi tube. The ETV can be designed more compact under the same signal intensity due to its significantly higher velocity in the throat section. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

17 pages, 5886 KiB

Open AccessEditor’s ChoiceArticle

An Estimation Method of Continuous Non-Invasive Arterial Blood Pressure Waveform Using Photoplethysmography: A U-Net Architecture-Based Approach

by Tasbiraha Athaya and Sunwoong Choi

Sensors 2021, 21(5), 1867; https://0-doi-org.brum.beds.ac.uk/10.3390/s21051867 - 07 Mar 2021

Cited by 66 | Viewed by 9702

Abstract

Blood pressure (BP) monitoring has significant importance in the treatment of hypertension and different cardiovascular health diseases. As photoplethysmogram (PPG) signals can be recorded non-invasively, research has been highly conducted to measure BP using PPG recently. In this paper, we propose a U-net [...] Read more.

Blood pressure (BP) monitoring has significant importance in the treatment of hypertension and different cardiovascular health diseases. As photoplethysmogram (PPG) signals can be recorded non-invasively, research has been highly conducted to measure BP using PPG recently. In this paper, we propose a U-net deep learning architecture that uses fingertip PPG signal as input to estimate arterial BP (ABP) waveform non-invasively. From this waveform, we have also measured systolic BP (SBP), diastolic BP (DBP), and mean arterial pressure (MAP). The proposed method was evaluated on a subset of 100 subjects from two publicly available databases: MIMIC and MIMIC-III. The predicted ABP waveforms correlated highly with the reference waveforms and we have obtained an average Pearson’s correlation coefficient of 0.993. The mean absolute error is 3.68 ± 4.42 mmHg for SBP, 1.97 ± 2.92 mmHg for DBP, and 2.17 ± 3.06 mmHg for MAP which satisfy the requirements of the Association for the Advancement of Medical Instrumentation (AAMI) standard and obtain grade A according to the British Hypertension Society (BHS) standard. The results show that the proposed method is an efficient process to estimate ABP waveform directly using fingertip PPG. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

27 pages, 2378 KiB

Open AccessArticle

EEG-Based Emotion Classification for Alzheimer’s Disease Patients Using Conventional Machine Learning and Recurrent Neural Network Models

by Jungryul Seo, Teemu H. Laine, Gyuhwan Oh and Kyung-Ah Sohn

Sensors 2020, 20(24), 7212; https://0-doi-org.brum.beds.ac.uk/10.3390/s20247212 - 16 Dec 2020

Cited by 14 | Viewed by 4267

Abstract

As the number of patients with Alzheimer’s disease (AD) increases, the effort needed to care for these patients increases as well. At the same time, advances in information and sensor technologies have reduced caring costs, providing a potential pathway for developing healthcare services [...] Read more.

As the number of patients with Alzheimer’s disease (AD) increases, the effort needed to care for these patients increases as well. At the same time, advances in information and sensor technologies have reduced caring costs, providing a potential pathway for developing healthcare services for AD patients. For instance, if a virtual reality (VR) system can provide emotion-adaptive content, the time that AD patients spend interacting with VR content is expected to be extended, allowing caregivers to focus on other tasks. As the first step towards this goal, in this study, we develop a classification model that detects AD patients’ emotions (e.g., happy, peaceful, or bored). We first collected electroencephalography (EEG) data from 30 Korean female AD patients who watched emotion-evoking videos at a medical rehabilitation center. We applied conventional machine learning algorithms, such as a multilayer perceptron (MLP) and support vector machine, along with deep learning models of recurrent neural network (RNN) architectures. The best performance was obtained from MLP, which achieved an average accuracy of 70.97%; the RNN model’s accuracy reached only 48.18%. Our study results open a new stream of research in the field of EEG-based emotion detection for patients with neurological disorders. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

19 pages, 3501 KiB

Open AccessArticle

Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model

by Yung-Hui Li, Latifa Nabila Harfiya, Kartika Purwandari and Yue-Der Lin

Sensors 2020, 20(19), 5606; https://0-doi-org.brum.beds.ac.uk/10.3390/s20195606 - 30 Sep 2020

Cited by 88 | Viewed by 6395

Abstract

Blood pressure monitoring is one avenue to monitor people’s health conditions. Early detection of abnormal blood pressure can help patients to get early treatment and reduce mortality associated with cardiovascular diseases. Therefore, it is very valuable to have a mechanism to perform real-time [...] Read more.

Blood pressure monitoring is one avenue to monitor people’s health conditions. Early detection of abnormal blood pressure can help patients to get early treatment and reduce mortality associated with cardiovascular diseases. Therefore, it is very valuable to have a mechanism to perform real-time monitoring for blood pressure changes in patients. In this paper, we propose deep learning regression models using an electrocardiogram (ECG) and photoplethysmogram (PPG) for the real-time estimation of systolic blood pressure (SBP) and diastolic blood pressure (DBP) values. We use a bidirectional layer of long short-term memory (LSTM) as the first layer and add a residual connection inside each of the following layers of the LSTMs. We also perform experiments to compare the performance between the traditional machine learning methods, another existing deep learning model, and the proposed deep learning models using the dataset of Physionet’s multiparameter intelligent monitoring in intensive care II (MIMIC II) as the source of ECG and PPG signals as well as the arterial blood pressure (ABP) signal. The results show that the proposed model outperforms the existing methods and is able to achieve accurate estimation which is promising in order to be applied in clinical practice effectively. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures

Figure 1

Other

Jump to: Research

14 pages, 914 KiB

Open AccessPerspective

A Sensor-Based Perspective in Early-Stage Parkinson’s Disease: Current State and the Need for Machine Learning Processes

by Marios G. Krokidis, Georgios N. Dimitrakopoulos, Aristidis G. Vrahatis, Christos Tzouvelekis, Dimitrios Drakoulis, Foteini Papavassileiou, Themis P. Exarchos and Panayiotis Vlamos

Sensors 2022, 22(2), 409; https://0-doi-org.brum.beds.ac.uk/10.3390/s22020409 - 06 Jan 2022

Cited by 13 | Viewed by 2755

Abstract

Parkinson’s disease (PD) is a progressive neurodegenerative disorder associated with dysfunction of dopaminergic neurons in the brain, lack of dopamine and the formation of abnormal Lewy body protein particles. PD is an idiopathic disease of the nervous system, characterized by motor and nonmotor [...] Read more.

Parkinson’s disease (PD) is a progressive neurodegenerative disorder associated with dysfunction of dopaminergic neurons in the brain, lack of dopamine and the formation of abnormal Lewy body protein particles. PD is an idiopathic disease of the nervous system, characterized by motor and nonmotor manifestations without a discrete onset of symptoms until a substantial loss of neurons has already occurred, enabling early diagnosis very challenging. Sensor-based platforms have gained much attention in clinical practice screening various biological signals simultaneously and allowing researchers to quickly receive a huge number of biomarkers for diagnostic and prognostic purposes. The integration of machine learning into medical systems provides the potential for optimization of data collection, disease prediction through classification of symptoms and can strongly support data-driven clinical decisions. This work attempts to examine some of the facts and current situation of sensor-based approaches in PD diagnosis and discusses ensemble techniques using sensor-based data for developing machine learning models for personalized risk prediction. Additionally, a biosensing platform combined with clinical data processing and appropriate software is proposed in order to implement a complete diagnostic system for PD monitoring. Full article

(This article belongs to the Special Issue Machine Learning for Sensing and Healthcare 2020–2021)

► Show Figures