A5016590855- Video Surveillance and Tracking Methods
- Vehicle License Plate Recognition
- Gait Recognition and Analysis
- Advanced Neural Network Applications
- Infrared Target Detection Methodologies
- Impact of Light on Environment and Health
- Face recognition and analysis
- Diabetes Management and Research
- AI in cancer detection
- Infrastructure Maintenance and Monitoring
- Brain Tumor Detection and Classification
- Handwritten Text Recognition Techniques
- Railway Engineering and Dynamics
- Nutrition and Health in Aging
- Mobile Health and mHealth Applications
- Diabetes and associated disorders
- Inflammatory Biomarkers in Disease Prognosis
- Cancer, Lipids, and Metabolism
- Radiomics and Machine Learning in Medical Imaging
- Biomedical Text Mining and Ontologies
- Autonomous Vehicle Technology and Safety
- Human Pose and Action Recognition
- Medical Image Segmentation Techniques
- Geotechnical Engineering and Underground Structures
- Railway Systems and Energy Efficiency
Beijing Information Science & Technology University
2024-2025
Beijing Shijitan Hospital
2023
Review Application of Wearable Devices in Diabetes Management Zijing Du 1,2,†, Feifan Zhang 1,†, Yifei Ge 1, Yijiang Liu 3, Honghua Yu 2, Yong Wang 4, Rinkoo Dalan 1,5, and Xiaotao Shen 1,3,* 1 Lee Kong Chian School Medicine, Nanyang Technological University, Singapore, 636921, Singapore 2 Guangdong Eye Institute, Department Ophthalmology, Provincial People’s Hospital (Guangdong Academy Medical Sciences), Southern Guangzhou, 510080, China 3 Chemistry, Chemical Engineering Biotechnology,...
In the field of medical image segmentation, models based on both CNN and Transformer have been thoroughly investigated. However, CNNs limited modeling capabilities for long-range dependencies, making it challenging to exploit semantic information within images fully. On other hand, quadratic computational complexity poses a challenge Transformers. State Space Models (SSMs), such as Mamba, recognized promising method. They not only demonstrate superior performance in interactions, but also...
Occluded person Re-identification (Re-ID) is to identify a particular when the person’s body parts are occluded. However, challenges remain in enhancing effective information representation and suppressing background clutter considering occlusion scenes. In this paper, we propose novel Attention Map-Driven Network (AMD-Net) for occluded Re-ID. AMD-Net, human parsing labels introduced supervise generation of partial attention maps, while suggest Spatial-frequency Interaction Module (SIM)...
The semantic segmentation task in pathology plays an indispensable role assisting physicians determining the condition of tissue lesions. With proposal Segment Anything Model (SAM), more and foundation models have seen rapid development field image segmentation. Recently, SAM2 has garnered widespread attention both natural medical Compared to SAM, it significantly improved terms accuracy generalization performance. We compared foundational based on SAM found that their performance...
Purpose The purpose of this study is to systematically investigate the novel phenomenon rail corrugation on small radius curves with joints in mountainous city metros, characterized by coexistence short and long wavelengths (30–40 mm 150–200 mm) low rail. Design/methodology/approach finite element model wheel-rail system section joint constructed based field surveys. friction-coupled vibration characteristics are studied from perspective friction self-excited feedback irregularity. Findings...
Current road damage detection methods, relying on manual inspections or sensor-mounted vehicles, are inefficient, limited in coverage, and often inaccurate, especially for minor damages, leading to delays safety hazards. To address these issues enhance real-time using street view image data (SVRDD), we propose DAPONet, a model incorporating three key modules: dual attention mechanism combining global local attention, multi-scale partial over-parameterization module, an efficient downsampling...
The development of deep learning has facilitated the application person re-identification (ReID) technology in intelligent security. Visible-infrared (VI-ReID) aims to match pedestrians across infrared and visible modality images enabling 24-hour surveillance. Current studies relying on unsupervised transformations as well inefficient embedding constraints bridge spectral differences between images, however, limit their potential performance. To tackle limitations above approaches, this...
Rationale: The incidence, progression, and prognosis of cancer can be affected by inflammation nutrition. Female patients have different inflammatory nutritional states depending on their age tumor types. It is important to screen for suitable prognostic indicators female populations.