A5023052125- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Context-Aware Activity Recognition Systems
- Image Enhancement Techniques
- Visual Attention and Saliency Detection
- IoT and Edge/Fog Computing
- Industrial Vision Systems and Defect Detection
- Video Surveillance and Tracking Methods
- Hand Gesture Recognition Systems
- Clostridium difficile and Clostridium perfringens research
- Human Pose and Action Recognition
- Data Management and Algorithms
- Domain Adaptation and Few-Shot Learning
- Data Mining Algorithms and Applications
- Advanced Image Fusion Techniques
- Web Data Mining and Analysis
- Smart Agriculture and AI
- Marine Biology and Environmental Chemistry
- Robot Manipulation and Learning
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Mobile Agent-Based Network Management
- Helicobacter pylori-related gastroenterology studies
- Microscopic Colitis
- Plant tissue culture and regeneration
National Institute for Communicable Disease Control and Prevention
2024-2025
Xi'an Polytechnic University
2023-2025
Henan Institute of Science and Technology
2016-2025
Center for Disease Control
2024
Northwestern Polytechnical University
2019-2024
Beijing Institute of Genomics
2024
Chinese Academy of Sciences
2016-2024
University of Jinan
2017-2023
Henan University
2021
East China Normal University
2021
Due to their rapid development and wide application in modern agriculture, robots, mobile terminals, intelligent devices have become vital technologies fundamental research topics for the of precision agriculture. Accurate efficient target detection technology is required inspection picking sorting equipment tomato production management plant factories. However, due limitations computer power, storage capacity, complexity factory (PF) environment, small-target tomatoes real-world...
The way people interact with machines through flexible acoustic sensors is revolutionizing the we live. However, developing a human-machine interaction sensor that simultaneously offers low cost, high stability, fidelity, and sensitivity remains significant challenge. In this study, based on sound-driven triboelectric nanogenerator was proposed. A poly(vinylidene fluoride) (PVDF)/graphene oxide (GO) composite nanofiber film obtained as dielectric layer electrospinning, copper-nickel alloy...
Clostridioides difficile is the predominant pathogen in hospital-acquired infections and antibiotic-associated diarrhea. Dedicated networks annual reports for C. surveillance have been established Europe North America, however extensive investigation on prevalence of infection (CDI) China limited. In this study, 1528 patients with diarrhea were recruited from seven geographically representative regions between July 2021 2022. The positivity rate toxigenic using real-time fluorescence...
Traditional displacement sensors and computer vision technologies exhibit significant limitations when measuring multiple targets, including restrictions related to the measurement environment, manual tagging wiring issues associated with sensors. This paper proposes a method for multi-point vibration that integrates bicubic interpolation, sparse optical flow oriented FAST rotated BRIEF (ORB) detection algorithms, addressing aforementioned limitations. The proposed eliminates need tags...
We treated a patient with refractory Clostridioides difficile infection through 42-day tapered and pulsed vancomycin regimen. In previous treatment courses, our had multiple recurrent infections. This case indicates that we should strengthen vigilance toward ST5 strains are positive for toxins A B binary cdtA cdtB.
Due to the existence of unfavorable factors such as turbid water quality and target occlusion, it is difficult obtain valid data features. repeated calculation similar data, real-time performance algorithm poor. In view above problems, this paper proposes a multi-AUV collaborative recognition method based on transfer-reinforcement learning. The features information which collected by are fused wavelet transformation affine invariance. similarity calculated Mahalanobis distance learning model...
Abstract The objective is to improve the recognition rate of white and near cotton‐coloured impurities in raw cotton against a single visible light source background. A lightweight detection network model without anchor boxes based on improved YOLO v4‐tiny proposed this paper weighted feature fusion (WFF). WFF strategy was used accuracy algorithm. Meanwhile, address disadvantage that obtained by K‐means algorithm clustering do not have global features, anchor‐free tiny decoupled head schemes...
Target tracking has always been a hot research topic in the field of computer vision. Tracking-Learning-Detection (TLD) is new algorithm for online learning proposed by Zdenek Kalal. In algorithm, computation consuming detection module relatively large. To solve this problem and improve we an method to adaptively update threshold variance classifier, which can effectively reduce number target boxes, real-time performance accuracy. Experiments have conducted compare improved TLD with original...
Complicated underwater environments, such as occlusion by foreign objects and dim light, cause serious loss of object features. This certainly increases the challenge accurate detection objects. To tackle above problems, we proposed an recognition algorithm combined with significant environmental First, salient features were extracted feature extraction network. Second, associated attention mechanism. Finally, contrast graph structure object's related was constructed. The model is compared...
By the acronym MOCOCO we refer to our view of ambient intelligence (AmI), pointing out three main characteristics: Mobility, Contextualization and Collaboration. The is a challenging research area focusing on ubiquitous computing, profiling practices, context awareness, human-centric computing interaction design. We are concretizing approach in platform called IMERA (French for Computer Augmented Environment Mobile Interaction). In order make work together several sensors, actuators mobile...
Compared with traditional imaging, the light field contains more comprehensive image information and higher quality. However, available data for reconstruction are limited, repeated calculation of seriously affects accuracy real-time performance multiperspective reconstruction. To solve problems, this paper proposes a method based on transfer reinforcement learning. Firstly, similarity measurement model is established. According to threshold source domain target domain, learning or feature...
Existing salient object detection networks are large, have many parameters, bulky and take up a lot of computational resources. Seriously hinder its application promotion in boning robot. To solve this problem, paper proposes lightweight saliency algorithm for real-time localization livestock meat bones. First, feature extraction network based on multi-scale attention is constructed the encoding stage. ensure that more adequate features extracted with fewer parameters. Second, fusion jump...
An efficient in vitro plant regeneration system via hypocotyl segments of tetraploid Isatis indigotica Fort. was established. Murashige and Skoog's (MS) Gamborg's (GB5) media were found to be superior White medium for promoting shoot regeneration. The highest (92 %) achieved from hypocotyls cultured on MS containing 8.9 μM benzyladenine (BA) 2.7 naphthaleneacetic acid (NAA), with an average 4.2 shoots developed per explant. Plant also improved when the explants basal 3 % (m/v) sucrose grown...
In order to meet the embedded application requirements of machine learning algorithm, intelligent target detection and recognition algorithm based on convolutional neural network corresponding optimal process are studied. Detailed structure analysis performance carried out. Based GPU platform, TensorRT technology is used accelerate including fp16 int8 inference modes. Satisfactory verification results achieved platform. addition, an integrated system real-time H.265 encoding decoding...
Target recognition is a fundamental research topic in the process of robot grasping. In this paper, we proposed an algorithm framework for object based on local naive Bayes nearest neighbor with Kinect. With emergence invariant feature detection, method features gradually becomes mainstream. Object realized through matching model current scene and models library property. Considering number may be as many dozens or even hundreds, I divide into coarse fine recognition, part adopts algorithm,...