A5103034166- Industrial Technology and Control Systems
- Advanced Sensor and Control Systems
- Cancer Research and Treatments
- Advanced Computational Techniques and Applications
- Advanced Algorithms and Applications
- Video Surveillance and Tracking Methods
- RNA and protein synthesis mechanisms
- Advanced Image and Video Retrieval Techniques
- Civil and Geotechnical Engineering Research
- RNA modifications and cancer
- Microbial Metabolism and Applications
- Geomechanics and Mining Engineering
- Industrial Vision Systems and Defect Detection
- Power Systems and Technologies
- Advanced Decision-Making Techniques
- Clostridium difficile and Clostridium perfringens research
- Advanced Neural Network Applications
- Remote Sensing and Land Use
- Soil, Finite Element Methods
- Viral gastroenteritis research and epidemiology
- Engineering Applied Research
- Image and Object Detection Techniques
- Grouting, Rheology, and Soil Mechanics
- Image Retrieval and Classification Techniques
- Engineering Structural Analysis Methods
Guangzhou University of Chinese Medicine
2024
Hangzhou Normal University
2024
Xi'an University of Science and Technology
2008-2023
Sun Yat-sen University
2023
Sun Yat-sen Memorial Hospital
2023
China Electronics Technology Group Corporation
2019-2020
Guangzhou Development Zone Hospital
2017-2018
General Hospital of Guangzhou Military Command
2008-2016
Specifica (United States)
2016
Xiaomi (China)
2016
Adenoma detection rate (ADR) is a key colonoscopy quality indicator in Western clinical literature. Our low ADR prompted us to assess novel methods improve performance. retrospective reports suggested that water exchange (WE) could increase ADR. However, most of these studies used pain score or intubation as the primary outcome. Here we test hypothesis WE significantly increases among Chinese colonoscopists and design prospective randomized controlled trial using our outcome.This...
Cloud computing is network-based computing, which enables people to run modeling applications on a virtual server or thousands of servers. With the development cloud technology, public can use power just like way they water, electricity, gas and telephone. Geographic information one resources over network, there are increasing demands spatial-intensive applications. The providers should take account geographic services. In this paper, we expound some key technologies such as Distributed File...
Abstract Background Oral immunization with vaccines may be an effective strategy for prevention of Clostridium difficile infection (CDI). However, application previously developed preventing CDI has been limited due to various reasons. Here, we a recombinant Lactococcus lactis oral vaccine and evaluated its effect on C. difficile- infected animal model established in golden hamsters attempt provide alternative prevention. Methods Recombinant L. was using the pTRKH2 plasmid, high-copy-number...
The emergence of nosocomial infections caused by hypervirulent and carbapenem-resistant
The aim of the present study was to identify risk genes in myocardial infarction. Microarray data GSE34198, containing from peripheral blood 49 infarction samples and 48 corresponding control samples, were downloaded Gene Expression Omnibus database screen differentially expressed (DEGs). DEGs used construct a protein‑protein interaction (PPI) network patient which feature identified using neighboring score method. recursive elimination (RFE) algorithm employed select among genes,...
Local Binary Pattern(LBP) is an operator to describe the local texture of face, which has characteristics rotation invariance and grayscale invariance. This paper proposed a face recognition arithmetic based on fusion layered LBP feature(F-LLBP), using extract feature, fusing each layer, reducing dimension with Principal Component Analysis Linear Discriminant Analysis(PCA-LDA) get set low dimensional characteristic are easy classify, finally KNN classify. The experiments ORL database...
Focusing on the problems of complex detection network structure, multiple parameters, and low accuracy in process traffic sign detection, a lightweight algorithm MC-YOLOv5 is proposed. Firstly, backbone feature extraction structure CSPDarkNet-53 YOLOv5 replaced by Mobilenetv2, depth separable convolution introduced to reduce parameters computation; Then, K-means++ clustering used generate priori anchor frame automatically; Finally, CBAM attention mechanism into strengthen representation...
BACKGROUND Cardiac device infection (CDI) is a serious complication of cardiovascular implantable electronic (CIED) implantations. Many risk factors have been identified, but several are still uncertain. This study aimed to identify and evaluate the factors. Moreover, an control protocol (ICP) was carried out, its efficacy in reducing CDIs investigated. MATERIAL AND METHODS A total 1259 patients who received permanent pacemaker (PPM) implantations were enrolled this 3-year period high-volume...
It is difficult to be accurately predicted for wind power generation's random, intermittent and volatility. According the strong chaotic characteristics of speed, optimal time delay embedding dimensions speed are determined by using a short-term prediction phase space reconstruction theory. After sample reconstructed, carried out BP neural network. The experimental results show that higher forecasting accuracy generation can obtained.
Apple leaf diseases have similar morphology and different spot sizes. Currently, in response to the low efficiency high missed detection rate of traditional artificial apple disease pest methods, an improved YOLOv7-tiny method is proposed. Firstly, ELEN module feature extraction network (Backone), DCNV3 (Deformable Convolutional Networks) introduced improve model accuracy; Secondly, using combination Space Depth (SPD) Separable Convolution (DSConv) perform SPD DSConv downsampling,...
Background subtraction is an important method in moving target detection. Its difficulty lies the establishment of dynamic background model. All kinds non-parametric density estimation methods are effective to solve this problem, but there obvious shadow on foreground image and can not clearly outline contour when moves forward backward, left right a short distance. In order K-Nearest Neighbor kernel based Gaussian filter proposed for used analyze probability pixels extract target....
With the continuous breakthroughs in semiconductor technology, more and fields such as production, manufacturing, consumption are becoming intelligent, production testing of IC ( integrated circuit) has become important. Research institute's ASIC(application-specific needs to trace data be tested under tri-temperature conditions. At present, domestic methods detecting by naked eye manual input package ID(identification) number separation high low temperature boxes from test system basically...
With the development of deep learning, performance vehicle detection algorithms based on learning is constantly improved, which plays an important role in construction intelligent transportation. Single-stage target model widely used real-time because its advantages speed. In view low rate small objects images, this article proposes a object method improved YOLOv4 algorithm, using k-means clustering algorithm to re-create anchors suitable for UA-Detrac dataset and improve PANet. Compared...