Guanzheng Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Guanzheng Liu has a total of 101 co-authors. Here are the top co-authors by number of shared publications:
- Changhong Wang
School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University, Shenzhen, Guangdong, China
11 shared publications - Keming Wei
School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China
7 shared publications Lianrong Zheng
Department of Biomedical Engineering, School of Engineering, Sun Yat-sen University and Key Laboratory of Sensing Technology and Biomedical Instrument of Guangdong Province, School of Engineering, Sun Yat-sen University, Guangzhou 510275, China and Guangdong Provincial Engineering and Technology Centre of Advanced and PorTable Medical Device, Guangzhou 510275, China
5 shared publicationsKunyang Li
Department of Biomedical Engineering, School of Engineering, Sun Yat-sen University, Guangzhou, 510275, China
4 shared publications- Mingjing Chen
Alfred E. Mann Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, United States
4 shared publications - Shan Wu
School of Biomedical Engineering, Sun Yat-Sen University, Shenzhen 518107, China
4 shared publications
Latest Publications
- Open Access
- 21 September 2026
- Published by MDPI AG in Preprints.org
Entropy Analysis of Electrocardiography and Heart Rate Variability in Heart Failure: A Methodology-Oriented Narrative Review and Illustrative Evidence Map
Entropy measures characterize complex dynamics in electrocardiography (ECG) and heart rate variability (HRV), but their interpretation depends on estimator, signal representation, and analytical configuration. This methodology-oriented narrative review organized 126 references by evidentiary function and applied role-appropriate interpretation boundaries. Within this evidence set, 53 human heart-failure (HF) or HF-relevant reports underwent structured mapping by method family (template, multiscale, symbolic/distributional, spectral/wavelet, or conditional/coupling), research task, data provenance, and validation design. Findings varied with configuration, rhythm, preprocessing, and recording length, while reused datasets or shared cohorts limited the independence of report-level evidence. Of 12 reports presenting diagnostic/classification performance, only two reported participant-separated internal evaluation, without establishing a fully nested leakage-free pipeline. A non-exclusive set of 10 reports examined clinical outcomes or clinically relevant event associations, all within single cohorts. Neither subset included independent external patient-cohort validation or identified calibration. Entropy should therefore be treated as a configuration-, representation-, rhythm-, and task-specific research tool for dynamic phenotyping or a candidate component of multivariable models, rather than a standalone clinical measure. Progress requires complete estimator reporting and design-appropriate participant separation; candidates intended for prediction or decision support additionally require appropriate comparators, external validation, calibration, and decision-relevant evaluation.
- 21 September 2026
- Published by MDPI AG in Preprints.org
Open AccessEntropy Analysis of Electrocardiography and Heart Rate Variability in Heart Failure: A Methodology-Oriented Narrative Review and Illustrative Evidence Map
Entropy measures characterize complex dynamics in electrocardiography (ECG) and heart rate variability (HRV), but their interpretation depends on estimator, signal representation, and analytical configuration. This methodology-oriented narrative review organized 126 references by evidentiary function and applied role-appropriate interpretation boundaries. Within this evidence set, 53 human heart-failure (HF) or HF-relevant reports underwent structured mapping by method family (template, multiscale, symbolic/distributional, spectral/wavelet, or conditional/coupling), research task, data provenance, and validation design. Findings varied with configuration, rhythm, preprocessing, and recording length, while reused datasets or shared cohorts limited the independence of report-level evidence. Of 12 reports presenting diagnostic/classification performance, only two reported participant-separated internal evaluation, without establishing a fully nested leakage-free pipeline. A non-exclusive set of 10 reports examined clinical outcomes or clinically relevant event associations, all within single cohorts. Neither subset included independent external patient-cohort validation or identified calibration. Entropy should therefore be treated as a configuration-, representation-, rhythm-, and task-specific research tool for dynamic phenotyping or a candidate component of multivariable models, rather than a standalone clinical measure. Progress requires complete estimator reporting and design-appropriate participant separation; candidates intended for prediction or decision support additionally require appropriate comparators, external validation, calibration, and decision-relevant evaluation.
- Open Access
- 29 May 2026
- Published by MDPI AG in Micromachines
- Vol. 17  (6) , 678
Physics-Informed Optimization for the Sub-Feature-Scale Fabrication of Hollow Microneedles via Digital Light Processing
To overcome low bioavailability and high trauma in inner ear therapies, targeted delivery across the round window membrane (RWM) via hollow microneedles (HMNs) offers a promising solution. However, the fabrication of high-aspect-ratio, small-size HMNs remains challenging. This study demonstrates the successful fabrication of small-outer-diameter HMNs using a 10 μm resolution digital light processing (DLP) system. Finite element analysis (FEA) identified a double tangent-arc transition as the optimal structural design for minimizing stress concentration. To manage the heightened parameter sensitivity at sub-feature-scale fabrication, a corrected curing index (CCI) model was established via a physics-informed regression approach incorporating polymerization kinetics and nonlinear spatial intensity distribution, achieving high fitting accuracy (R2 > 0.96). Under optimized parameters, the fabricated HMNs possessed mean dimensions of 805.13 μm in height, 37.54 μm in inner diameter, and 79.36 μm in outer diameter. Compressive tests exhibited a robust structural strength of up to 141 mN per needle following post-curing. Combined in silico and in vitro experiments demonstrated excellent penetration performance. Furthermore, the HMNs achieved stable, pressure-dependent delivery with volumetric flow rates rising from 0.14 mL∙min−1 to 0.39 mL∙min−1 as driving pressure escalated from 50 kPa to 300 kPa, validating their functional capacity for controlled drug administration.
- 29 May 2026
- Published by MDPI AG in Micromachines
- Vol. 17  (6) , 678
Open AccessPhysics-Informed Optimization for the Sub-Feature-Scale Fabrication of Hollow Microneedles via Digital Light Processing
To overcome low bioavailability and high trauma in inner ear therapies, targeted delivery across the round window membrane (RWM) via hollow microneedles (HMNs) offers a promising solution. However, the fabrication of high-aspect-ratio, small-size HMNs remains challenging. This study demonstrates the successful fabrication of small-outer-diameter HMNs using a 10 μm resolution digital light processing (DLP) system. Finite element analysis (FEA) identified a double tangent-arc transition as the optimal structural design for minimizing stress concentration. To manage the heightened parameter sensitivity at sub-feature-scale fabrication, a corrected curing index (CCI) model was established via a physics-informed regression approach incorporating polymerization kinetics and nonlinear spatial intensity distribution, achieving high fitting accuracy (R2 > 0.96). Under optimized parameters, the fabricated HMNs possessed mean dimensions of 805.13 μm in height, 37.54 μm in inner diameter, and 79.36 μm in outer diameter. Compressive tests exhibited a robust structural strength of up to 141 mN per needle following post-curing. Combined in silico and in vitro experiments demonstrated excellent penetration performance. Furthermore, the HMNs achieved stable, pressure-dependent delivery with volumetric flow rates rising from 0.14 mL∙min−1 to 0.39 mL∙min−1 as driving pressure escalated from 50 kPa to 300 kPa, validating their functional capacity for controlled drug administration.
- 30 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71  (4) , 9820-9831
An Ordinal Pattern Similarity-Guided Supervised Prototype Contrastive Learning Framework With Enhanced Token Selection Module for Sleep Apnea Detection Based on Wearable PPG Bracelet
Sleep apnea (SA) is a chronic condition with wide-ranging effects, yet screening remains limited. SA detection using wearable smart wristband devices has the potential to significantly improve this situation. However, existing detection methods suffer from limited representation capacity and achieve low sensitivity. To this end, we propose an ordinal pattern similarity-guided supervised prototype contrastive learning (ops-spcl) framework with enhanced token selection module. Specifically, in the self-attention mechanism of the token selection module, we integrate a selective kernel convolution-ReLU operation. By incorporating receptive fields of varying scales, this approach enhances the model’s ability to focus more precisely on target-relevant token features, thereby improving the capture of temporal dependencies in SA events. Furthermore, considering the challenges of accurately capturing the features of SA events in the peak-to-peak interval sequences of mild to moderate patients, we design an ops-spcl loss function. This function explicitly constructs hard negative sample pairs and easy positive sample pairs to enhance intra-class compactness and inter-class separability that generates a more refined, effective, and discriminative feature space. On the wearable bracelet PPG dataset, the proposed model performs well in per-segment detection, achieving 72.4% sensitivity and a 0.687 F1-score. By analyzing SA detection rates across different population groups, we demonstrate that it effectively focuses on the hard samples of mild to moderate patients. Moreover, these findings highlight the generality of its strong generalization capabilities across multiple datasets. These results suggest that our approach holds significant potential for large-scale SA detection. Our code is publicly available.
- 30 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71  (4) , 9820-9831
An Ordinal Pattern Similarity-Guided Supervised Prototype Contrastive Learning Framework With Enhanced Token Selection Module for Sleep Apnea Detection Based on Wearable PPG Bracelet
Sleep apnea (SA) is a chronic condition with wide-ranging effects, yet screening remains limited. SA detection using wearable smart wristband devices has the potential to significantly improve this situation. However, existing detection methods suffer from limited representation capacity and achieve low sensitivity. To this end, we propose an ordinal pattern similarity-guided supervised prototype contrastive learning (ops-spcl) framework with enhanced token selection module. Specifically, in the self-attention mechanism of the token selection module, we integrate a selective kernel convolution-ReLU operation. By incorporating receptive fields of varying scales, this approach enhances the model’s ability to focus more precisely on target-relevant token features, thereby improving the capture of temporal dependencies in SA events. Furthermore, considering the challenges of accurately capturing the features of SA events in the peak-to-peak interval sequences of mild to moderate patients, we design an ops-spcl loss function. This function explicitly constructs hard negative sample pairs and easy positive sample pairs to enhance intra-class compactness and inter-class separability that generates a more refined, effective, and discriminative feature space. On the wearable bracelet PPG dataset, the proposed model performs well in per-segment detection, achieving 72.4% sensitivity and a 0.687 F1-score. By analyzing SA detection rates across different population groups, we demonstrate that it effectively focuses on the hard samples of mild to moderate patients. Moreover, these findings highlight the generality of its strong generalization capabilities across multiple datasets. These results suggest that our approach holds significant potential for large-scale SA detection. Our code is publicly available.
- 13 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29  (11) , 8011-8021
An EEG Screening Method for Severe Obstructive Sleep Apnea Based on Limited Penetrable Difference Visibility Graph and Graph Convolutional Network
Obstructive Sleep Apnea (OSA) is a common respiratory disease characterized by recurrent airway block during sleep, which does great harm to the human body. Utilizing the electroencephalogram (EEG) has been proven instrumental in OSA detection, as sleep apnea occurrences induce discernible alterations in EEG patterns. In this study, we propose a Limited Penetrable Difference Visibility Graph (LPDVG) method to screen severe OSA. This method exhibits strong anti-noise performance, effective information extraction, and a certain degree of generalization ability. First of all, this study constructed LPDVG complex network and calculated the information entropy of the degree sequence in six leads. Subsequently, this study weighted the information entropy of each lead using the mutual information between leads to fuse information from the whole brain. Eventually, a classification model for the Graph Convolutional Network (GCN) was trained to detect patients with severe OSA. Using a dataset of 88 participants, we tested and evaluated this approach. The results showed a strong correlation between the extracted feature and AHI, with a Pearson correlation of 0.792. The accuracy, specificity, sensitivity, and area under the curve (AUC) of the GCN classification were 82.95%, 83.87%, 80.77%, and 0.905. Moreover, there are significant differences in LPDVGwSEN between patients in severe and non-severe OSA groups. Compared to those with non-severe OSA, brain activity in patients with severe OSA is more disorganized, especially in the theta frequency band. EEG data indicating this elevated activity is associated with disturbed sleep patterns in individuals with OSA.
- 13 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29  (11) , 8011-8021
An EEG Screening Method for Severe Obstructive Sleep Apnea Based on Limited Penetrable Difference Visibility Graph and Graph Convolutional Network
Obstructive Sleep Apnea (OSA) is a common respiratory disease characterized by recurrent airway block during sleep, which does great harm to the human body. Utilizing the electroencephalogram (EEG) has been proven instrumental in OSA detection, as sleep apnea occurrences induce discernible alterations in EEG patterns. In this study, we propose a Limited Penetrable Difference Visibility Graph (LPDVG) method to screen severe OSA. This method exhibits strong anti-noise performance, effective information extraction, and a certain degree of generalization ability. First of all, this study constructed LPDVG complex network and calculated the information entropy of the degree sequence in six leads. Subsequently, this study weighted the information entropy of each lead using the mutual information between leads to fuse information from the whole brain. Eventually, a classification model for the Graph Convolutional Network (GCN) was trained to detect patients with severe OSA. Using a dataset of 88 participants, we tested and evaluated this approach. The results showed a strong correlation between the extracted feature and AHI, with a Pearson correlation of 0.792. The accuracy, specificity, sensitivity, and area under the curve (AUC) of the GCN classification were 82.95%, 83.87%, 80.77%, and 0.905. Moreover, there are significant differences in LPDVGwSEN between patients in severe and non-severe OSA groups. Compared to those with non-severe OSA, brain activity in patients with severe OSA is more disorganized, especially in the theta frequency band. EEG data indicating this elevated activity is associated with disturbed sleep patterns in individuals with OSA.
- 13 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74  (00189456) , 1-10
Graph-Weighted High-Dimensional Information-Based Similarity for Detecting Altered Brain Functional Connectivity in Sleep Apnea Patients
Sleep apnea (SA) is a common respiratory disease associated with changes in the cortical connectivity pattern of the brain, leading to a growing social and medical burden. The electroencephalogram (EEG)-based analysis can help to understand the pathological mechanism of SA in the brain. Information-based similarity (IBS) is an effective method for measuring the nonlinear similarity of two physiological signals and has the potential for SA screening. However, the traditional IBS method can only extract local information, failing to fully capture pathological EEG patterns characterized by critical cross-frequency dynamics and regional variability. To address these issues, we propose graph-weighted high-dimensional information-based similarity (HGIBS), which integrates 2-D IBS into global information, thus improving the accuracy and reliability of SA detection. We use EEG energy sequences from multiple brain regions and frequency bands, divide their wave patterns into several modes, and calculate their similarity by comparing the distribution of the appearance frequency of the modes. Additionally, we construct the brain network graph and extend the calculation of similarity to high-dimensional space, thereby integrating local information into global information. We evaluated the performance of the HGIBS method on dataset of 88 participants and achieved a correlation between HGIBS and apnea-hypopnea index (AHI) of 0.730, with an accuracy of 89.4% in severe SA classification, demonstrating the superiority. This method has the potential to become a new SA screening indicator and provide a deeper understanding of the pathological mechanism of SA research.
- 13 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74  (00189456) , 1-10
Graph-Weighted High-Dimensional Information-Based Similarity for Detecting Altered Brain Functional Connectivity in Sleep Apnea Patients
Sleep apnea (SA) is a common respiratory disease associated with changes in the cortical connectivity pattern of the brain, leading to a growing social and medical burden. The electroencephalogram (EEG)-based analysis can help to understand the pathological mechanism of SA in the brain. Information-based similarity (IBS) is an effective method for measuring the nonlinear similarity of two physiological signals and has the potential for SA screening. However, the traditional IBS method can only extract local information, failing to fully capture pathological EEG patterns characterized by critical cross-frequency dynamics and regional variability. To address these issues, we propose graph-weighted high-dimensional information-based similarity (HGIBS), which integrates 2-D IBS into global information, thus improving the accuracy and reliability of SA detection. We use EEG energy sequences from multiple brain regions and frequency bands, divide their wave patterns into several modes, and calculate their similarity by comparing the distribution of the appearance frequency of the modes. Additionally, we construct the brain network graph and extend the calculation of similarity to high-dimensional space, thereby integrating local information into global information. We evaluated the performance of the HGIBS method on dataset of 88 participants and achieved a correlation between HGIBS and apnea-hypopnea index (AHI) of 0.730, with an accuracy of 89.4% in severe SA classification, demonstrating the superiority. This method has the potential to become a new SA screening indicator and provide a deeper understanding of the pathological mechanism of SA research.
- 25 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12  (14) , 25989-26002
DIGS-Module: A Deformable Information Gain-Guided Scale Selection Structure Based on Physiological Signals for Classification
Multiscale convolutional neural networks are widely utilized in the analysis of physiological signals to assess human health due to their large receptive field and rich feature domain. However, current scale selection for multiscale networks relies heavily on data and empirical knowledge, leading to information redundancy and reduced generalization capability, which hinders clinical application. To address these, we propose a novel paradigm of deformable information gain-guided scale (DIGS), which adaptively selects the appropriate scale for feature extraction according to the input signal. Specifically, we employ an information entropy-based hierarchical calculation of the information gain between the input and output features of each multiscale branch. This information gain determines the optimal branches for feature extraction. Additionally, we utilize 1-D deformable module to achieve overall convolutional shifts, enabling scale information to match the distribution of the dataset. Extensive experiments conducted on five datasets demonstrate that our proposed multiscale paradigm outperforms previous baseline networks in four classification tasks of physiological signals. Furthermore, integrating our structure with the best performing network on current private wearable PPG datasets improves accuracy by approximately 1%, sensitivity by 1%–3%, and reduces computational load by nearly half. These findings indicate that our new structure enhances both classification performance and computational efficiency of multiscale networks.
- 25 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 12  (14) , 25989-26002
DIGS-Module: A Deformable Information Gain-Guided Scale Selection Structure Based on Physiological Signals for Classification
Multiscale convolutional neural networks are widely utilized in the analysis of physiological signals to assess human health due to their large receptive field and rich feature domain. However, current scale selection for multiscale networks relies heavily on data and empirical knowledge, leading to information redundancy and reduced generalization capability, which hinders clinical application. To address these, we propose a novel paradigm of deformable information gain-guided scale (DIGS), which adaptively selects the appropriate scale for feature extraction according to the input signal. Specifically, we employ an information entropy-based hierarchical calculation of the information gain between the input and output features of each multiscale branch. This information gain determines the optimal branches for feature extraction. Additionally, we utilize 1-D deformable module to achieve overall convolutional shifts, enabling scale information to match the distribution of the dataset. Extensive experiments conducted on five datasets demonstrate that our proposed multiscale paradigm outperforms previous baseline networks in four classification tasks of physiological signals. Furthermore, integrating our structure with the best performing network on current private wearable PPG datasets improves accuracy by approximately 1%, sensitivity by 1%–3%, and reduces computational load by nearly half. These findings indicate that our new structure enhances both classification performance and computational efficiency of multiscale networks.
- 5 February 2025
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 105
Multi-task self-supervised learning framework via contrastive restoration for sleep apnea syndrome detection using single-lead ECG
- 5 February 2025
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 105
Multi-task self-supervised learning framework via contrastive restoration for sleep apnea syndrome detection using single-lead ECG
- 30 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 16  (2) , 783-798
A Topic-Guided Self-Attention Network for Daily Mental Wellbeing Prediction Using Mobile Devices
Prediction of daily mental wellbeing holds profound implications for individual healthcare and societal stability. Previous studies have shown the potential of using individual's multimodal behavioral data collected through mobile devices to predict his/her daily mental wellbeing metrics, such as stress, mood, and anxiety. However, effectively capturing long-range dependencies in behavioral time series data while accurately representing the statistical distribution patterns of various behaviors over a certain period is a significant challenge. In this paper, we propose a daily mental wellbeing prediction model based on a Topic-Guided Self-Attention Network (TGSAN). This model utilizes self-attention mechanism to capture long-range dependencies from the behavioral data collected by mobile devices. We utilize a multi-granularity time encoding method to inject time information of different granularities (i.e., day and hour, or week and day) into the behavioral data, thereby enhancing the sensibility of the self-attention network to capture every individual's habitual cyclicality rhythm. Then, we introduce a neural topic model to analyze the statistical distribution characteristics of various behaviors in the monitoring period as behavioral distribution patterns for different individuals, and further propose a topic attention network to enhance the model's classification performance by guiding the weights of long-range dependencies features from the self-attention network with the derived topic information. Compared to state-of-the-art methods, the proposed TGSAN achieved superior performance on datasets that measure different mental health indicators (stress, mood, and anxiety), with F1 scores outperforming by 4.5% and 2.3% on the Crosscheck and StudentLife datasets, respectively, and accuracy outperforming by 3.3% on the GLOBEM dataset. Our study demonstrates the effectiveness and interpretability of combining self-attention mechanisms with neural topic model, for a better understanding of the relationship between different individuals' behaviors and their mental wellbeing.
- 30 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 16  (2) , 783-798
A Topic-Guided Self-Attention Network for Daily Mental Wellbeing Prediction Using Mobile Devices
Prediction of daily mental wellbeing holds profound implications for individual healthcare and societal stability. Previous studies have shown the potential of using individual's multimodal behavioral data collected through mobile devices to predict his/her daily mental wellbeing metrics, such as stress, mood, and anxiety. However, effectively capturing long-range dependencies in behavioral time series data while accurately representing the statistical distribution patterns of various behaviors over a certain period is a significant challenge. In this paper, we propose a daily mental wellbeing prediction model based on a Topic-Guided Self-Attention Network (TGSAN). This model utilizes self-attention mechanism to capture long-range dependencies from the behavioral data collected by mobile devices. We utilize a multi-granularity time encoding method to inject time information of different granularities (i.e., day and hour, or week and day) into the behavioral data, thereby enhancing the sensibility of the self-attention network to capture every individual's habitual cyclicality rhythm. Then, we introduce a neural topic model to analyze the statistical distribution characteristics of various behaviors in the monitoring period as behavioral distribution patterns for different individuals, and further propose a topic attention network to enhance the model's classification performance by guiding the weights of long-range dependencies features from the self-attention network with the derived topic information. Compared to state-of-the-art methods, the proposed TGSAN achieved superior performance on datasets that measure different mental health indicators (stress, mood, and anxiety), with F1 scores outperforming by 4.5% and 2.3% on the Crosscheck and StudentLife datasets, respectively, and accuracy outperforming by 3.3% on the GLOBEM dataset. Our study demonstrates the effectiveness and interpretability of combining self-attention mechanisms with neural topic model, for a better understanding of the relationship between different individuals' behaviors and their mental wellbeing.
- 19 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25  (1) , 737-748
A Novel Heart Rate Variability Measurement Method for Detecting Obstructive Sleep Apnea Using Local Difference Visibility Graphs
Obstructive sleep apnea (OSA) is a common and chronic respiratory disorder. Autonomic dysfunction and heart rate change usually occur during apneic events. Damaged autonomic nerve system modulation can lead to morphological differences in R-to-R intervals (RRI) series. Capable of capturing structural properties of RRI series and thus assessing the heart rate dynamics, the local difference visibility graphs (LDVG) method was proposed as a novel OSA detection method in this study. The LDVG method focuses on long continuous ascent range, which is likely to be induced by OSA attack, demonstrating the characteristics of bradycardia and tachycardia in OSA. Combined with the local strategy, it avoids heavy workload while maintaining major correlations. Degree mean of LDVG (LDVG_M) and degree uncertainty of LDVG (LDVG_SEn) were calculated as the mean and Shannon entropy of the degree sequence of LDVG, which can quantify the tightness and uncertainty of structural characteristics of the RR interval series, respectively. 60 ECG recordings from the Physionet database were used in this study. Both LDVG_M and LDVG_SEn achieved a good performance in differentiating the severity of diseases, while the classic index low-frequency power to high-frequency power ratio value (LF/HF) cannot. Then, binary classification was performed. In OSA detection, the accuracy, specificity, sensitivity of both LDVG-related indices were all above 90%. LDVG_SEn achieved the highest accuracy (96.7%), while maintaining a good balance between sensitivity (97.5%) and specificity (95.0%). Therefore, this LDVG method provides an insight into the underlying pathological mechanism of OSA and is promising for clinical and home monitoring.
- 19 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25  (1) , 737-748
A Novel Heart Rate Variability Measurement Method for Detecting Obstructive Sleep Apnea Using Local Difference Visibility Graphs
Obstructive sleep apnea (OSA) is a common and chronic respiratory disorder. Autonomic dysfunction and heart rate change usually occur during apneic events. Damaged autonomic nerve system modulation can lead to morphological differences in R-to-R intervals (RRI) series. Capable of capturing structural properties of RRI series and thus assessing the heart rate dynamics, the local difference visibility graphs (LDVG) method was proposed as a novel OSA detection method in this study. The LDVG method focuses on long continuous ascent range, which is likely to be induced by OSA attack, demonstrating the characteristics of bradycardia and tachycardia in OSA. Combined with the local strategy, it avoids heavy workload while maintaining major correlations. Degree mean of LDVG (LDVG_M) and degree uncertainty of LDVG (LDVG_SEn) were calculated as the mean and Shannon entropy of the degree sequence of LDVG, which can quantify the tightness and uncertainty of structural characteristics of the RR interval series, respectively. 60 ECG recordings from the Physionet database were used in this study. Both LDVG_M and LDVG_SEn achieved a good performance in differentiating the severity of diseases, while the classic index low-frequency power to high-frequency power ratio value (LF/HF) cannot. Then, binary classification was performed. In OSA detection, the accuracy, specificity, sensitivity of both LDVG-related indices were all above 90%. LDVG_SEn achieved the highest accuracy (96.7%), while maintaining a good balance between sensitivity (97.5%) and specificity (95.0%). Therefore, this LDVG method provides an insight into the underlying pathological mechanism of OSA and is promising for clinical and home monitoring.
- 10 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 11  (16) , 27139-27152
A Multiscale Cross-Modal Interactive Fusion Network for Human Activity Recognition Using Wearable Sensors and Smartphones
Human activity recognition (HAR) enables real-time monitoring of human movement, posture, and activity level, and can provide valuable information for the health management. With the continuous advancement of Internet of Things (IoT) technology, wearable sensors and smartphones equipped with various types of sensors have become widely utilized to collect multimodal data for HAR. However, in multimodal HAR, current fusion methods fall short in capturing intermodality correlations, hampering the full exploitation of complementary information between the modalities and leading to lower recognition accuracy. We thus propose a novel multiscale cross-modal interactive fusion network (MCIFN), which can fully capture correlations between various modalities and obtain an effective fused representation for HAR. Specifically, we employ a multiscale parallel convolution module to extract features from each modality at multiple scales. Then, an interactive fusion strategy based on the cross-modal attention mechanism is introduced to adjust and enhance each modality based on its correlations with other modalities. Additionally, to resolve the information redundancy caused by the interactive fusion strategy, we utilize a hybrid attention module to focus on important information in the fusion representation. Extensive experiments conducted on three publicly available data sets and one private data set demonstrate that our proposed network outperforms the previous baseline networks for HAR. Additionally, our proposed fusion strategy yielded a notable improvement in accuracy ranging from 1.87% to 9.96% compared to existing strategies. These findings imply that our newly proposed network can realize comprehensive multimodal fusion and effectively enhance HAR accuracy, potentially contributing to advancements in individual health management and personalized healthcare interventions.
- 10 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 11  (16) , 27139-27152
A Multiscale Cross-Modal Interactive Fusion Network for Human Activity Recognition Using Wearable Sensors and Smartphones
Human activity recognition (HAR) enables real-time monitoring of human movement, posture, and activity level, and can provide valuable information for the health management. With the continuous advancement of Internet of Things (IoT) technology, wearable sensors and smartphones equipped with various types of sensors have become widely utilized to collect multimodal data for HAR. However, in multimodal HAR, current fusion methods fall short in capturing intermodality correlations, hampering the full exploitation of complementary information between the modalities and leading to lower recognition accuracy. We thus propose a novel multiscale cross-modal interactive fusion network (MCIFN), which can fully capture correlations between various modalities and obtain an effective fused representation for HAR. Specifically, we employ a multiscale parallel convolution module to extract features from each modality at multiple scales. Then, an interactive fusion strategy based on the cross-modal attention mechanism is introduced to adjust and enhance each modality based on its correlations with other modalities. Additionally, to resolve the information redundancy caused by the interactive fusion strategy, we utilize a hybrid attention module to focus on important information in the fusion representation. Extensive experiments conducted on three publicly available data sets and one private data set demonstrate that our proposed network outperforms the previous baseline networks for HAR. Additionally, our proposed fusion strategy yielded a notable improvement in accuracy ranging from 1.87% to 9.96% compared to existing strategies. These findings imply that our newly proposed network can realize comprehensive multimodal fusion and effectively enhance HAR accuracy, potentially contributing to advancements in individual health management and personalized healthcare interventions.