A5101523244- Energy, Environment, Economic Growth
- Energy and Environment Impacts
- Advanced Neural Network Applications
- Microfluidic and Bio-sensing Technologies
- Biometric Identification and Security
- Neuroinflammation and Neurodegeneration Mechanisms
- 3D Printing in Biomedical Research
- Computer Graphics and Visualization Techniques
- Environmental Impact and Sustainability
- Face and Expression Recognition
- Microfluidic and Capillary Electrophoresis Applications
- Advanced Vision and Imaging
- Generative Adversarial Networks and Image Synthesis
- Face recognition and analysis
- AI in cancer detection
- Gout, Hyperuricemia, Uric Acid
- Digital Media Forensic Detection
- Metal Extraction and Bioleaching
- Cervical Cancer and HPV Research
- Mine drainage and remediation techniques
- Machine Learning and Data Classification
- Domain Adaptation and Few-Shot Learning
- Network Security and Intrusion Detection
- Advanced biosensing and bioanalysis techniques
- Electromagnetic wave absorption materials
Beijing Institute of Technology
2021-2025
Northern Jiangsu People's Hospital
2023-2024
Yangzhou University
2023-2024
Xinjiang University
2022-2024
Beijing University of Chemical Technology
2022-2023
University of Science and Technology Beijing
2022
Guangdong University of Technology
2021
South China University of Technology
2015-2020
Beijing University of Technology
2019
University of Otago
2015
Xanthine oxidase (XO), a form of xanthine oxidoreductase, is widely distributed in various human tissues. As major source for the generation superoxide radicals, XO involved induction oxidative stress and inflammation during ischemic hypoxic tissue injury. Therefore, we designed this study to identify role serum levels acute stroke (AIS) pathogenesis. In single-center prospective study, 328 consecutive patients with AIS first time were included, 107 age- sex-matched healthy controls from...
Face recognition technique has been widely applied to personal identification systems due its satisfying performance. However, security may be a crucial issue, since many studies have shown that face vulnerable in an adversarial environment, which adversary can camouflage as legitimate user order mislead the system. Although liveness detection methods proposed distinguish real and fake faces, they are either time-consuming, costly, or sensitive noise illumination. This paper proposes method...
Pseudomonas syringae pv. actinidiae is an economically significant pathogen responsible for severe bacterial canker of kiwifruit (Actinidia sp.). Bacteriophages infecting this phytopathogen have potential as biocontrol agents part integrated approach to the management canker, and use molecular tools study bacterium. A variety bacteriophages were previously isolated that infect P. actinidiae, their basic properties characterized provide a framework formulation these phages agents. Here, we...
Based on the demand for a volatile organic compounds (VOCs) emissions inventory Beijing and reduction in VOC emissions, three major solvent-based industries of automobile manufacturing, furniture publication printing were selected this study. In each case, link negative pressure sampling was used combination with laboratory-based GC-MS/FID to obtain spectra. The results show that there are significant differences emission VOCs from main stages manufacturing. Specifically, paint-coating...
Automation‐assisted cervical screening via liquid‐based cytology has achieved great success using segmentation and classification methods. This work tries to do abnormal region detection on field of view cell images based deep learning, which is a novel way solve cytological problem. Since some nuclei gather in groups, the proposed method chooses regions instead as targets order locate for further diagnosis pathologists. In this study, approach size‐sensitive fully convolutional network...
The traditional classification technique cannot be applied directly to the big data problem mainly due high time complexity issue. existing method trained a classifier using dataset simplified by clustering approach. However, main drawback of this is that may over- or under-simplified. In paper, we revises for task. Our proposed considers multi-view rather than single training set. Several datasets are firstly generated based on with different radii. Then several classifiers these and...
Chinese flower-bird characters are gems of traditional art. It is an artistic font and the strokes designed as beautiful patterns flowers, birds, etc. The generation such requires painters' great efforts. Imagine that if we only need to sketch outline ideal using a pencil, then can quickly obtain these characters, which will be significance in promoting their development, allowing more people appreciate even create this art by computer. Recently, with development deep learning invention...
For real-time semantic segmentation, how to expedite while maintaining the resolution of high quality is an ongoing and unresolved problem.The backbone design has always been two important parts segmentation. We designed a lightweight through multiple reuse features achieve most advanced In order this goal, we use encoder-decoder architecture solve problem. encoder new model downsampling methods for segmentation task. decoder have three different ways fuse semantics detailed information...
Small object detection is widely used in real-world scenarios. It a hot topic the field of detection. However, because small objects occupy fewer pixels and are easily disturbed by background, their results far inferior to those large medium-sized objects. Therefore, we propose feature enhancement network based on YOLOv5 (FE-YOLOv5) improve performance We modify original layers design module (FEM) spatial-aware (SAM). The FEM guides shallow features using global attention high-level context...
With the accelerated development of technological power society, aerial images drones gradually penetrated various industries. Due to variable speed drones, captured are shadowed, blurred, and obscured. Second, fly at varying altitudes, leading changing target scales making it difficult detect identify small targets. In order solve above problems, an improved ASG-YOLOv5 model is proposed in this paper. Firstly, research proposes a dynamic contextual attention module, which uses feature...