A5101515037- Gait Recognition and Analysis
- Topic Modeling
- Diabetic Foot Ulcer Assessment and Management
- Muscle activation and electromyography studies
- Non-Invasive Vital Sign Monitoring
- Natural Language Processing Techniques
- Anomaly Detection Techniques and Applications
- Cardiovascular Health and Disease Prevention
- Balance, Gait, and Falls Prevention
- Hand Gesture Recognition Systems
- Blood Pressure and Hypertension Studies
- Context-Aware Activity Recognition Systems
- Hemodynamic Monitoring and Therapy
- Multimodal Machine Learning Applications
- Simulation and Modeling Applications
- Advanced Sensor and Energy Harvesting Materials
- Text and Document Classification Technologies
- ECG Monitoring and Analysis
- Advanced Image and Video Retrieval Techniques
- Digital Media Forensic Detection
- Human Pose and Action Recognition
- Botulinum Toxin and Related Neurological Disorders
- Heart Rate Variability and Autonomic Control
- Cardiovascular Health and Risk Factors
- Advanced MEMS and NEMS Technologies
Chinese Academy of Sciences
2016-2025
University of Science and Technology of China
2010-2025
Hefei Institutes of Physical Science
2018-2025
Johns Hopkins University
2025
Chengdu Second People's Hospital
2025
Jilin Normal University
2024
Institute of Intelligent Machines
2014-2023
China State Shipbuilding (China)
2023
McGill University
2023
King's College London
2021
Early detection of arrhythmia and effective treatment can prevent deaths caused by cardiovascular disease (CVD). In clinical practice, the diagnosis is made checking electrocardiogram (ECG) beat-by-beat, but this usually time-consuming laborious. paper, we propose an automatic ECG classification method based on Continuous Wavelet Transform (CWT) Convolutional Neural Network (CNN). CWT used to decompose signals obtain different time-frequency components, CNN extract features from 2D-scalogram...
A facile deprotection-free method to in situ grow GDY on a C 3 N 4 surface is developed. This straightforward strategy enables the formation of an integral GDY@C heterojunction, leading enhanced photocatalytic activity.
Freezing of gait (FOG) is a paroxysmal dyskinesia, which common in patients with advanced Parkinson’s disease (PD). It an important cause falls PD and associated serious disability. In this study, we implemented novel FOG detection system using deep learning technology. The takes multi-channel acceleration signals as input, uses one-dimensional convolutional neural network to automatically learn feature representations, recurrent model the temporal dependencies between activations. order...
Hypertension is a widespread chronic disease. Risk prediction of hypertension an intervention that contributes to the early prevention and management hypertension. The implementation such requires effective easy-to-implement risk model. This study evaluated compared performance four machine learning algorithms on predicting based easy-to-collect factors. A dataset 29,700 samples collected through physical examination was used for model training testing. Firstly, we identified factors...
In this work, chitosan (CS) decorated metronidazole (MTZ) microcapsules (CS@MTZ) were synthesized and used as a cross-linker for the preparation of poly(vinyl alcohol) (PVA) injectable hydrogel by dynamic covalent bonding ionic interaction through 4-carboxyphenylboronic acid bridge. The use MTZ efficiently slowed down release rate hydrophilic antibiotic from matrix. Besides, hydrophobicity endows PVA@CS@MTZ to be sticky substrate in wet conditions, under suggested mechanism evicting water...
This study proposed a muscle fatigue classification method based on surface electromyography (sEMG) signals to achieve accurate detection and classification. A total of 20 healthy young participants (14 men 6 women) were recruited for testing cycle ergometer, sEMG oxygen uptake recorded during the test. First, measured denoised with an improved wavelet threshold method. Second, V-slope was used identify ventilation (VT) reflect state. The time- frequency-domain features extracted, including...
Human activity recognition (HAR) is essential in many health-related fields. A variety of technologies based on different sensors have been developed for HAR. Among them, fusion from heterogeneous wearable has as it portable, non-interventional and accurate To be applied real-time use with limited resources, the system must compact reliable. This requirement can achieved by feature selection (FS). By eliminating irrelevant redundant features, burden reduced good classification performance...
Previous studies have used the anaerobic threshold (AT) to non-invasively predict muscle fatigue. This study proposes a novel method for automatic classification of fatigue based on surface electromyography (sEMG). The sEMG data were acquired from 20 participants during an incremental test cycle ergometer using sensors placed vastus rectus femoris (RF), lateralis (VL), medialis (VM), and gastrocnemius (GA) muscles left leg. ventilation volume (VE), oxygen uptake (VO2), carbon dioxide...
The medical literature contains valuable knowledge, such as the clinical symptoms, diagnosis, and treatments of a particular disease. Named Entity Recognition (NER) is initial step in extracting this knowledge from unstructured text presenting it Knowledge Graph (KG). However, previous approaches NER have often suffered small-scale human-labelled training data. Furthermore, Chinese more complex task because there no segmentation between characters. Recently, pretraining models, which obtain...
ABSTRACT To ensure precise and rapid identification of casting surface defects to support the subsequent realisation high‐precision grinding, this study introduces a method for detecting using lightweight YOLOv5 framework. The enhanced model integrates ShuffleNetV2 high‐efficiency CNN architecture into foundation, substantially reducing network parameters achieve model. Additionally, Convolutional Block Attention Module (CBAM) attention mechanism is incorporated enhance model's capability...
Recent researches of large language models(LLM), which is pre-trained on massive general-purpose corpora, have achieved breakthroughs in responding human queries. However, these methods face challenges including limited data insufficiency to support extensive pre-training and can not align responses with users' instructions. To address issues, we introduce a medical instruction dataset, CMedINS, containing six instructions derived from actual tasks, effectively fine-tunes LLM conjunction...
With the rapid development of 3D reconstruction technology, widespread distribution data has become a future trend. While traditional visual (such as images and videos) NeRF-based formats already have mature techniques for copyright protection, steganographic emerging Gaussian Splatting (3D-GS) format yet to be fully explored. To address this, we propose ConcealGS, an innovative method embedding implicit information into 3D-GS. By introducing knowledge distillation gradient optimization...
This study aims to investigate nursing staff's current knowledge-attitude-practice(KAP) regarding document quality control and explore effective methods enhance their awareness of documents' importance through intensive training. We developed the questionnaire based on a systematic literature review two rounds Delphi expert consultation. Then, we sent nurses before after training.Data processing statistical analysis were conducted using R 4.4.0 software. Altogether, 722, 701, 800...
Abstract Background Sleep is essential for normal and healthy living. Lack of good quality sleep affects physical, mental emotional functions. Currently, the treatments obesity-related disorders focus more on suppressing sleep-related symptoms pharmaceutically are often accompanied by side effects. Thus, there urgent need alternative ways to combat chronic disorders. This study will investigate underlying mechanisms effects exercise diet intervention disorders, role gut microbiota in...
Intelligent transportation systems (ITS) can improve the efficiency and safety of transportation. Vehicular Ad Hoc Networks (VANETs) are an important foundation ITS. However, before being deployed on a large scale, VANETs should resolve security privacy issues generated from wireless communication. Digital signature has been used in to construct anonymous authentication schemes which realize preservation. Many previous have preloaded master private key into tamper-proof devices (TPD)...
Named Entity Recognition (NER) is the fundamental task for Natural Language Processing (NLP) and initial step in building a Knowledge Graph (KG). Recently, BERT (Bidirectional Encoder Representations from Transformers), which pre-training model, has achieved state-of-the-art (SOTA) results various NLP tasks, including NER. However, Chinese NER still more challenging because there are no physical separations between words, can only obtain representations of characters. Nevertheless, cannot be...
Named Entity Recognition (NER) is a fundamental Natural Language Processing (NLP) task to extract entities from unstructured data. The previous methods for NER were based on machine learning or deep learning. Recently, pre-training models have significantly improved performance multiple NLP tasks. In this paper, firstly, we introduce the architecture and tasks of four common models: BERT, ERNIE, ERNIE2.0-tiny, RoBERTa. Then, apply these by fine-tuning, compare effects different model task....
Among the ever-evolving development of vision-language models, contrastive language-image pretraining (CLIP) has set new benchmarks in many downstream tasks such as zero-shot classifications by leveraging self-supervised learning on large amounts text-image pairs. However, its dependency rigid one-to-one mappings overlooks complex and often multifaceted relationships between within texts images. To this end, we introduce RankCLIP, a novel method that extends beyond matching framework CLIP...