A5100767773- Advanced Neural Network Applications
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Data Management and Algorithms
- Data Visualization and Analytics
- Multimodal Machine Learning Applications
- Infrared Target Detection Methodologies
- Geographic Information Systems Studies
- Time Series Analysis and Forecasting
- Big Data Technologies and Applications
- Fiscal Policy and Economic Growth
- Advanced Image Processing Techniques
- Advanced biosensing and bioanalysis techniques
- Industrial Vision Systems and Defect Detection
- Consumer Market Behavior and Pricing
- Topic Modeling
- Digital Transformation in Industry
- Digital Marketing and Social Media
- Energy, Environment, Economic Growth
- Visual Attention and Saliency Detection
- Consumer Retail Behavior Studies
- Local Government Finance and Decentralization
- Supramolecular Self-Assembly in Materials
- Regional Development and Environment
- Numerical Methods and Algorithms
Capital University of Economics and Business
2024-2025
Zhengzhou University of Light Industry
2024
Beijing University of Chemical Technology
2023-2024
Macau University of Science and Technology
2024
Qilu University of Technology
2022-2023
Shandong Academy of Sciences
2022-2023
Tianjin University of Technology
2022-2023
Shenzhen University
2022
Soochow University
2022
China University of Mining and Technology
2019-2021
Traditional camera sensors rely on human eyes for observation. However, are prone to fatigue when observing objects of different sizes a long time in complex scenes, and cognition is limited, which often leads judgment errors greatly reduces efficiency. Object recognition technology an important used judge the object’s category sensor. In order solve this problem, small-size object detection algorithm special scenarios was proposed paper. The advantage that it not only has higher precision...
Traditional camera sensors rely on human eyes for observation. However, the eye 1
 is prone to fatigue when observing targets of different sizes a long time in complex scenes, and 2
 cognition limited, which often leads judgment errors greatly reduces efficiency. 3
 Target recognition technology an important judge target category 4
 sensor. In order solve this problem, small size detection algorithm special scenarios was 5
 proposed by paper. Its advantage that not...
YOLOv5 is one of the most popular object detection algorithms, which divided into multiple series according to control network depth and width. To realize deployment mobile devices or embedded devices, paper proposes a lightweight aerial image algorithm (LAI-YOLOv5s) based on improvement YOLOv5s with relatively small amount calculation parameter fast reasoning speed. Firstly, better detect objects, replaces minimum head maximum new feature fusion method, DFM-CPFN(Deep Feature Map Cross Path...
Driven by deep learning, great breakthroughs had been made in the field of target detection. Small detection algorithms were widely used industry, agriculture and other fields. But small few available features loss detail information feature extraction. So it led to low accuracy algorithms. In this paper, we proposed DBF‐YOLO algorithm based on classical YOLOV5. The YOLOV5 with high speed. speed minimum model could reach 24 ms. However, network structure targets. Our was an improvement...
As the road traffic situation becomes complex, task of management takes on an increasingly heavy load. The air-to-ground administration network drones has become important tool to promote high quality police work in many places. Drones can be used instead a large number human beings perform daily tasks, as: offense detection, crowd etc. are aerial operations and shoot small targets. So detection accuracy is less. To address problem low Unmanned Aerial Vehicles (UAVs) detecting targets, we...
Using CsOH as the only catalyst and utilizing its “cesium effect”, a clean synthesis of wide range primary, secondary, tertiary carboxamides was achieved by aminolysis reactions nitriles with ammonia, or secondary amines in water.
The "Three Rural Issues", encompass challenges related to agriculture, farmer, and rural area, which hold significant importance in driving comprehensive revitalization efforts China. Farmer entrepreneurship, as a crucial means enhance productivity, create job opportunities, increase residents’ income, has gradually become key force promoting the new stage of development With rapid e-commerce, farmer entrepreneurship encountered opportunities. This study utilizes 2020 China Family Panel...
An in-depth understanding of the impact green innovation on urban–rural income gap is essential for developing countries seeking to address imbalances and promote sustainable economic development. This study focuses China, utilizing provincial panel data from 2007 2022, employs two-way fixed effects model, mediating moderating model. The study’s key findings are: (1) Green positively reduces gap. Specifically, regression results indicate that a one-unit increase in corresponds 0.017-unit...
Abstract Reusable Launch Vehicles are crucial for reducing space mission costs and improving efficiency through multiple flights. The unpowered return automatic landing phase pose challenges due to aerodynamic uncertainties, the absence of propulsion, strict constraints. This research integrates Gauss pseudospectral method with a three-degree-of-freedom simulation model plan trajectory. includes Earth’s curvature, ground effects, gear deployment. A Linear Active Disturbance Rejection...
The detection of multi-class small objects poses a significant challenge in the field computer vision. While original YOLOv5 algorithm is more suited for detecting full-scale objects, it may not perform optimally this specific task. To address issue, we proposed MC-YOLOv5, an specifically designed object detection. Our approach incorporates three key innovations: (1) application improved CB module during feature extraction to capture edge information that be less apparent thereby enhancing...
A robust fiscal and financial system constitutes a fundamental pillar of national governance. This paper investigates the spatial correlation between decentralization, local green economic efficiency using panel data from 285 cities in China. The findings reveal several key insights. First, decentralization hampers enhancement but exerts “warning effect” on neighboring regions. In contrast, promotes triggers “clustering areas, indicating an agglomeration impact. Second, synergies are not...
How to obtain discriminative features has proved be a core problem for image retrieval. Many recent works use convolutional neural networks extract features. However, clutter and occlusion will interfere with the distinguishability of when using network (CNN) feature extraction. To address this problem, we intend high-response activations in map based on attention mechanism. We propose two modules, spatial module channel module. For module, first capture global information model relation...
With the global outbreak of Corona Virus Disease 2019(COVID-19), many countries had made it mandatory for people to wear masks in public places. This paper proposed a novel mask detection algorithm RMPC (Restructing Maxpool layer and Convolution layer)-YOLOv7 based on YOLOv7 detecting whether The RMPC-YOLOv7 reconstructed downsampling structure original algorithm. We changed stacking maxpooling convolutional layer. enabled feature information be fully integrated achieve accuracy improvement...
Person Re-Identification (re-id) is a challenging task in computer vision, especially when there are limited training data from multiple camera views. In this paper, we pro- pose deep learning based person re-identification method by transferring knowledge of mid-level attribute features and high-level classification features. Building on the idea that identity classification, recognition re- identification share same semantic representations, they can be trained sequentially fine-tuning one...
As a frontier research topic in the field of scientific visualization, trajectory data visualization extracts valuable patterns and knowledge from for decision support via spatiotemporal techniques. We propose concept multivariate interpret two categories attributes that are based on geographical space abstract space. Properly analyzing depends many factors such as task sparsity. Therefore, we generalize rich interactions to explore evolution events transform issue into more intelligibly...
The Coronavirus Disease 2019 (COVID-19) has proved a globally prevalent outbreak since December 2019. As focused country to alleviate the epidemic impact, China implemented range of public health interventions prevent disease from further transmission, including pandemic lockdown in Wuhan and other cities. This paper establishes China’s mobility network by flight dataset proposes model without epidemiological parameters indicate spread risks through network, which is termed as strength. By...
Abstract Now deep learn-based object detection can be deployed on drones for criminalinvestigation or military counter-terrorism. Because the proportion of pixels pedestrians orvehicles in aerial picture taken by UAV is very small, probability smallobjects distance low there are omissions. In this paper, HPS-YOLOv7 algorithmis proposed to improve accuracy small objects. We have a modifiedhigh-efficiency layer aggregation network feature extraction, solved problem that theconvergence depth...
Facing the ever-increasing demands of microprocessors for computing power nowadays, trigonometric functions, as basis various algorithms, have problems with latency, resource overhead and consumption in traditional software or coprocessor schemes. In this paper, a lightweight RISC-V standard microprocessor is designed. addition, optimization CORDIC algorithm realizes low-latency high-precision trigonometrical kernel-level accelerator, instruction set extended. The experiment shows that...
The purpose of network security is to prevent the data transmitted over Internet from being stolen and tampered with, ensure data. It not only necessary that information entering exiting or but also integrity confidentiality in system. environment becoming more complex, attack methods are diverse. Therefore, intrusion detection systems have some common problems, such as low rate high false alarm rate, it difficult meet real-time requirements systems. Currently, deep learning increasingly...
Abstract Relationship and trust are essential factors to aggregate the agricultural product supply chain’s central bodies maintain dynamic flexibility. Especially for chain dominated by e-commerce with many atomic farmers, relationships even more critical. The basic premise ensure its effective operation. Big data blockchain has high security, distributed storage, publicity mechanism, anonymity, openness, transparency, etc. It is of great significance apply it products. Based on analysis...
This paper reviews the NTIRE 2024 challenge on image super-resolution ($\times$4), highlighting solutions proposed and outcomes obtained. The involves generating corresponding high-resolution (HR) images, magnified by a factor of four, from low-resolution (LR) inputs using prior information. LR images originate bicubic downsampling degradation. aim is to obtain designs/solutions with most advanced SR performance, no constraints computational resources (e.g., model size FLOPs) or training...