Changhong Wang
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
Changhong Wang has a total of 77 co-authors. Here are the top co-authors by number of shared publications:
- Guanzheng Liu
School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China
11 shared publications - Bijan Najafi
Center for Advanced Surgical & Interventional Technology (CASIT), Department of Surgery, David Geffen School of Medicine, University of California, Los Angeles (UCLA), Los Angeles, CA 90095, USA
5 shared publications - Stephen James Redmond
School of Electrical and Electronic Engineering, University College Dublin, Dublin, Ireland
4 shared publications - Yanxun Lu
School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China
4 shared publications - Zeju Xu
School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China
4 shared publications - Haodong Liu
School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China
3 shared publications
Latest Publications
- Open Access
- 23 July 2026
- Published by Springer Nature in Nature Communications
- Vol. 17  (1)
A programmable living force sensor fabricated with surface engineered microbial network
Engineered living materials have recently expanded into the field of bioelectronics when incorporating the rationally rewired microbes inside. Integrating living materials with force sensing provides an alternative route to overcome limitations of traditional sensors, such as fragility and stretching-sensitivity. Here, we present a programmable living force sensor fabricated by surface engineered microbial network with a bottom-up design for human-machine interfaces. Bacterial cells are genetically rewired to display surface-anchored affinity pairs that mediate the assembly of bacteria to macroscale. Functioning as the conductive and sensing layer, the microbial network is encapsulated in elastic tubes to fabricate the macroscopic force sensor. The resulting living sensor features a tunable measuring range, rapid self-recovery capacity within tens of milliseconds, and remarkably, excellent stretching-insensitivity with only ~30% increase in resistance under 800% of elongation. This bio-integrated approach enables microbes to prepare robust and reconfigurable force sensor suitable for tensile, dynamic and extreme mechanical conditions.
- 23 July 2026
- Published by Springer Nature in Nature Communications
- Vol. 17  (1)
Open AccessA programmable living force sensor fabricated with surface engineered microbial network
Engineered living materials have recently expanded into the field of bioelectronics when incorporating the rationally rewired microbes inside. Integrating living materials with force sensing provides an alternative route to overcome limitations of traditional sensors, such as fragility and stretching-sensitivity. Here, we present a programmable living force sensor fabricated by surface engineered microbial network with a bottom-up design for human-machine interfaces. Bacterial cells are genetically rewired to display surface-anchored affinity pairs that mediate the assembly of bacteria to macroscale. Functioning as the conductive and sensing layer, the microbial network is encapsulated in elastic tubes to fabricate the macroscopic force sensor. The resulting living sensor features a tunable measuring range, rapid self-recovery capacity within tens of milliseconds, and remarkably, excellent stretching-insensitivity with only ~30% increase in resistance under 800% of elongation. This bio-integrated approach enables microbes to prepare robust and reconfigurable force sensor suitable for tensile, dynamic and extreme mechanical conditions.
- Open Access
- 11 June 2026
- Published by Springer Nature in npj Digital Medicine
- Vol. 9  (1)
Daily activity patterns from wearable accelerometry predict physical frailty and concern about falling
Physical frailty and concern about falling are interrelated geriatric conditions that significantly impair mobility, limit daily activity, and reduce life quality among older adults. Traditional methods for assessing frailty and fall concern rely heavily on questionnaires such as the Fried Frailty Phenotype (FFP) and the Falls Efficacy Scale-International (FES-I). These tools, although clinically established, are subjective, require professional administration, and cannot capture the mutual influence between the two conditions. In this study, we propose a wearable-sensor-based, objective framework that simultaneously predicts physical frailty status and level of concern about falling by analyzing real-world activity patterns. We collected continuous chest acceleration data from 146 participants over a 48-hour period using a pendant sensor. These activity sequences were transformed into barcodes and analyzed using global complexity metrics (e.g., single- and multi-scale entropy) and local sequential dynamics via bidirectional LSTM networks. A multi-task deep learning model with attention mechanisms was proposed to jointly predict frailty and fall concern. Our model achieved high predictive performance for both tasks, achieving F1 scores of 93.12% for physical frailty and 86.27% for concern about falling. This study provides the first joint modeling of physical frailty and concern about falling using long-term real-world accelerometry data.
- 11 June 2026
- Published by Springer Nature in npj Digital Medicine
- Vol. 9  (1)
Open AccessDaily activity patterns from wearable accelerometry predict physical frailty and concern about falling
Physical frailty and concern about falling are interrelated geriatric conditions that significantly impair mobility, limit daily activity, and reduce life quality among older adults. Traditional methods for assessing frailty and fall concern rely heavily on questionnaires such as the Fried Frailty Phenotype (FFP) and the Falls Efficacy Scale-International (FES-I). These tools, although clinically established, are subjective, require professional administration, and cannot capture the mutual influence between the two conditions. In this study, we propose a wearable-sensor-based, objective framework that simultaneously predicts physical frailty status and level of concern about falling by analyzing real-world activity patterns. We collected continuous chest acceleration data from 146 participants over a 48-hour period using a pendant sensor. These activity sequences were transformed into barcodes and analyzed using global complexity metrics (e.g., single- and multi-scale entropy) and local sequential dynamics via bidirectional LSTM networks. A multi-task deep learning model with attention mechanisms was proposed to jointly predict frailty and fall concern. Our model achieved high predictive performance for both tasks, achieving F1 scores of 93.12% for physical frailty and 86.27% for concern about falling. This study provides the first joint modeling of physical frailty and concern about falling using long-term real-world accelerometry data.
- 4 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. PP  (1530437X) , 1
Efficient Behavioral Context Recognition with Multi-Label Early-Exit and Dynamic Multimodal Aggregation on Edge Devices
Behavioral context recognition aims to identify an individual’s situational state and has broad application prospects across multiple fields. With the proliferation of smart devices such as smartphones and wearables, sensor data have become powerful tools for recognizing human behavior. Although deep neural networks have shown strong performance in behavioral context recognition, they typically rely on static inference structures. However, given the inherently multimodal and multi-label nature of this task, such static designs exacerbate computational complexity, leading to excessive computational demands that hinder deployment on resource-constrained devices. To address these issues, we propose EBCR, a computation-efficient framework for multimodal multi-label behavioral context recognition. The framework employs a hierarchical modality aggregation mechanism to reduce redundant multimodal computation and introduces a relation-aware loss function to explicitly model label dependencies, thereby improving both recognition accuracy and the reliability of early-exit decisions. On the Extrasensory and ETRI Lifelong 2020 datasets, EBCR achieves a balanced accuracy of 74.6% and 63.0%, respectively, comparable to state-of-the- art methods. When deployed on a Raspberry Pi 5, EBCR operates with only 19.21–25.95 MFLOPs, achieving 4–5 ms latency and 0.020–0.032 J energy consumption, resulting in 1.1×–2.5× faster inference and 17.6%–66.7% lower energy cost compared to existing models. These results demonstrate the effectiveness of EBCR in achieving accurate and efficient multi-label behavioral context recognition under real-world resource constraints.
- 4 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. PP  (1530437X) , 1
Efficient Behavioral Context Recognition with Multi-Label Early-Exit and Dynamic Multimodal Aggregation on Edge Devices
Behavioral context recognition aims to identify an individual’s situational state and has broad application prospects across multiple fields. With the proliferation of smart devices such as smartphones and wearables, sensor data have become powerful tools for recognizing human behavior. Although deep neural networks have shown strong performance in behavioral context recognition, they typically rely on static inference structures. However, given the inherently multimodal and multi-label nature of this task, such static designs exacerbate computational complexity, leading to excessive computational demands that hinder deployment on resource-constrained devices. To address these issues, we propose EBCR, a computation-efficient framework for multimodal multi-label behavioral context recognition. The framework employs a hierarchical modality aggregation mechanism to reduce redundant multimodal computation and introduces a relation-aware loss function to explicitly model label dependencies, thereby improving both recognition accuracy and the reliability of early-exit decisions. On the Extrasensory and ETRI Lifelong 2020 datasets, EBCR achieves a balanced accuracy of 74.6% and 63.0%, respectively, comparable to state-of-the- art methods. When deployed on a Raspberry Pi 5, EBCR operates with only 19.21–25.95 MFLOPs, achieving 4–5 ms latency and 0.020–0.032 J energy consumption, resulting in 1.1×–2.5× faster inference and 17.6%–66.7% lower energy cost compared to existing models. These results demonstrate the effectiveness of EBCR in achieving accurate and efficient multi-label behavioral context recognition under real-world resource constraints.
- 16 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. PP, 1-15
Heterogeneous Multimodal Multilabel Behavioral Context Recognition Using Wearable Sensors and Mobile Devices
Accurate behavioral context recognition (BCR) is essential for advancing social intelligence in smart home settings and social companion pairing. With advancements in mobile computing technologies, wearable sensors, and mobile devices have become crucial for multimodal data collection to interpret behavioral context. However, existing methods may overlook the multilabel attribute of behavioral context, limiting their ability to model modality-to-label and label-to-label dependencies. Moreover, the unique label heterogeneity in BCR, arising from category-specific differences among labels, is often underestimated. To overcome these challenges, we propose a heterogeneous multimodal multilabel recognition method to capture the relationships between modalities and labels, while taking the label heterogeneity into account. Using specialized heterogeneous decoders with self-adaptive attention and a heterogeneous graph attention network based on the dual-level attention mechanism, our method captures complex dependencies among labels and modalities. Experiments on two benchmark datasets demonstrate the superior performance of our approach, with ablation studies validating the contributions of each component.
- 16 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. PP, 1-15
Heterogeneous Multimodal Multilabel Behavioral Context Recognition Using Wearable Sensors and Mobile Devices
Accurate behavioral context recognition (BCR) is essential for advancing social intelligence in smart home settings and social companion pairing. With advancements in mobile computing technologies, wearable sensors, and mobile devices have become crucial for multimodal data collection to interpret behavioral context. However, existing methods may overlook the multilabel attribute of behavioral context, limiting their ability to model modality-to-label and label-to-label dependencies. Moreover, the unique label heterogeneity in BCR, arising from category-specific differences among labels, is often underestimated. To overcome these challenges, we propose a heterogeneous multimodal multilabel recognition method to capture the relationships between modalities and labels, while taking the label heterogeneity into account. Using specialized heterogeneous decoders with self-adaptive attention and a heterogeneous graph attention network based on the dual-level attention mechanism, our method captures complex dependencies among labels and modalities. Experiments on two benchmark datasets demonstrate the superior performance of our approach, with ablation studies validating the contributions of each component.
- 13 March 2026
- Published by Elsevier in Pattern Recognition
- Vol. 178
Semi-supervised topic-guided contrastive learning with data augmentation for daily stress prediction using smartphone sensors
- 13 March 2026
- Published by Elsevier in Pattern Recognition
- Vol. 178
Semi-supervised topic-guided contrastive learning with data augmentation for daily stress prediction using smartphone sensors
- 17 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11  (2) , 2186-2193
Ultra-Fast Lightweight Incipient Slip Detection Using Hyperdimensional Computing With the PapillArray Tactile Sensor
Timely detection of incipient slip is critical for delicate robotic grasping and dexterous manipulation. However, existing learning-based methods suffer from detection latency and high computational demands. In this paper, we present an ultra-fast lightweight incipient slip detection framework based on hyperdimensional (HD) computing, using the PapillArray optical tactile sensor. Our approach introduces a novel graphical-spatial-temporal HD encoding scheme coupled with a context-driven training and inference strategy, achieving a slip detection accuracy of 91.78% in offline evaluation. The resulting model is exceptionally compact and highly edge-compatible, with a size of only 0.375 kB. Furthermore, hardware acceleration on an FPGA enables inference within 0.42 microseconds, representing an over $10^{4}\,\times$ speedup compared to optimized CPU implementations. Online robotic experiments involving grip-force control based on the proposed slip detection method further validate its practical effectiveness. This work offers a practical and scalable solution for real-time slip detection in robotic manipulation tasks.
- 17 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11  (2) , 2186-2193
Ultra-Fast Lightweight Incipient Slip Detection Using Hyperdimensional Computing With the PapillArray Tactile Sensor
Timely detection of incipient slip is critical for delicate robotic grasping and dexterous manipulation. However, existing learning-based methods suffer from detection latency and high computational demands. In this paper, we present an ultra-fast lightweight incipient slip detection framework based on hyperdimensional (HD) computing, using the PapillArray optical tactile sensor. Our approach introduces a novel graphical-spatial-temporal HD encoding scheme coupled with a context-driven training and inference strategy, achieving a slip detection accuracy of 91.78% in offline evaluation. The resulting model is exceptionally compact and highly edge-compatible, with a size of only 0.375 kB. Furthermore, hardware acceleration on an FPGA enables inference within 0.42 microseconds, representing an over $10^{4}\,\times$ speedup compared to optimized CPU implementations. Online robotic experiments involving grip-force control based on the proposed slip detection method further validate its practical effectiveness. This work offers a practical and scalable solution for real-time slip detection in robotic manipulation tasks.
- 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.
- Open Access
BiHDTrans: binary hyperdimensional transformer for efficient multivariate time series classification
The proliferation of Internet-of-Things (IoT) devices has led to an unprecedented volume of multivariate time series (MTS) data, requiring efficient and accurate processing for timely decision-making in resource-constrained edge environments. Hyperdimensional (HD) computing, with its inherent efficiency and parallelizability, has shown promise in classification tasks but struggles to capture complex temporal patterns, while Transformers excel at sequence modeling but incur high computational and memory overhead. We introduce BiHDTrans, an efficient neurosymbolic binary hyperdimensional Transformer that integrates self-attention into the HD computing paradigm, unifying the representational efficiency of HD computing with the temporal modeling power of Transformers. Empirically, BiHDTrans outperforms state-of-the-art (SOTA) HD computing models by at least 14.47% and achieves 6.67% higher accuracy on average than SOTA binary Transformers. With hardware acceleration on FPGA, our pipelined implementation leverages the independent and identically distributed properties of high-dimensional representations, delivering 39.4 times lower inference latency than SOTA binary Transformers. Theoretical analysis shows that binarizing in holographic high-dimensional space incurs significantly less information distortion than directly binarizing neural networks, explaining BiHDTrans's superior accuracy. Furthermore, dimensionality experiments confirm that BiHDTrans remains competitive even with a 64% reduction in hyperspace dimensionality, surpassing SOTA binary Transformers by 1-2% in accuracy with 4.4 times less model size, as well as further reducing the latency by 49.8% compare to the full-dimensional baseline. Together, these contributions bridge the gap between the expressiveness of Transformers and the efficiency of HD computing, enabling accurate, scalable, and low-latency MTS classification.
Open AccessBiHDTrans: binary hyperdimensional transformer for efficient multivariate time series classification
The proliferation of Internet-of-Things (IoT) devices has led to an unprecedented volume of multivariate time series (MTS) data, requiring efficient and accurate processing for timely decision-making in resource-constrained edge environments. Hyperdimensional (HD) computing, with its inherent efficiency and parallelizability, has shown promise in classification tasks but struggles to capture complex temporal patterns, while Transformers excel at sequence modeling but incur high computational and memory overhead. We introduce BiHDTrans, an efficient neurosymbolic binary hyperdimensional Transformer that integrates self-attention into the HD computing paradigm, unifying the representational efficiency of HD computing with the temporal modeling power of Transformers. Empirically, BiHDTrans outperforms state-of-the-art (SOTA) HD computing models by at least 14.47% and achieves 6.67% higher accuracy on average than SOTA binary Transformers. With hardware acceleration on FPGA, our pipelined implementation leverages the independent and identically distributed properties of high-dimensional representations, delivering 39.4 times lower inference latency than SOTA binary Transformers. Theoretical analysis shows that binarizing in holographic high-dimensional space incurs significantly less information distortion than directly binarizing neural networks, explaining BiHDTrans's superior accuracy. Furthermore, dimensionality experiments confirm that BiHDTrans remains competitive even with a 64% reduction in hyperspace dimensionality, surpassing SOTA binary Transformers by 1-2% in accuracy with 4.4 times less model size, as well as further reducing the latency by 49.8% compare to the full-dimensional baseline. Together, these contributions bridge the gap between the expressiveness of Transformers and the efficiency of HD computing, enabling accurate, scalable, and low-latency MTS classification.
- Open Access
- 21 August 2025
- Published by Elsevier in JACC: Advances
- Vol. 4  (11) , 101993
Automatic Diagnosis of Left Valvular Heart Disease Based on Artificial Intelligence Stethoscope
- 21 August 2025
- Published by Elsevier in JACC: Advances
- Vol. 4  (11) , 101993
Open AccessAutomatic Diagnosis of Left Valvular Heart Disease Based on Artificial Intelligence Stethoscope
- 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.