A5101733974- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Remote-Sensing Image Classification
- Infrared Target Detection Methodologies
- Quantum Computing Algorithms and Architecture
- Automated Road and Building Extraction
- Adversarial Robustness in Machine Learning
- Satellite Image Processing and Photogrammetry
- Quantum Information and Cryptography
- Deception detection and forensic psychology
- Domain Adaptation and Few-Shot Learning
- Inertial Sensor and Navigation
- Neural Networks and Reservoir Computing
- Edible Oils Quality and Analysis
- Multimodal Machine Learning Applications
- Optical Systems and Laser Technology
- Spam and Phishing Detection
- Data Mining Algorithms and Applications
- Anomaly Detection Techniques and Applications
- Web Data Mining and Analysis
- Data Visualization and Analytics
- Fatty Acid Research and Health
- Privacy-Preserving Technologies in Data
- Lipid metabolism and biosynthesis
Hunan University
2021-2025
Tianjin University
2025
China Centre for Resources Satellite Data and Application
2024
Zhejiang Lab
2023
Tianjin University of Technology and Education
2022
Beijing University of Chemical Technology
2019
Beihang University
2018-2019
Chinese People's Armed Police Force Engineering University
2016
Institute of Agro-Products Processing Science and Technology
2008-2012
Chinese Academy of Agricultural Sciences
2008-2012
Ship detection is an important and challenging task in remote sensing applications. Most methods utilize specially designed hand-crafted features to detect ships, they usually work well only on one scale, which lack generalization impractical identify ships with various scales from multiresolution images. In this paper, we propose a novel deep feature-based method very high-resolution optical our method, regional proposal network used generate ship candidates feature maps produced by...
Vehicle detection is a significant and challenging task in aerial remote sensing applications. Most existing methods detect vehicles with regular rectangle boxes fail to offer the orientation of vehicles. However, information crucial for several practical applications, such as trajectory motion estimation In this paper, we propose novel deep network, called rotatable region-based residual network (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Abstract Gaussian boson sampling (GBS) has the potential to solve complex graph problems, such as clique finding, which is relevant drug discovery tasks. However, realizing full benefits of quantum enhancements requires large-scale hardware with universal programmability. Here we have developed a time-bin-encoded GBS photonic processor that universal, programmable and software-scalable. Our features freely adjustable squeezing parameters can implement arbitrary unitary operations...
The intake of edible oil containing trans-fatty acids has deleterious effects mainly on the cardiovascular system. Thermal processes such as refining and frying cause formation in oil. This study was conducted to investigate possible because heat treatment soybean types heated are determined by attenuated total reflectance Fourier transform infrared spectroscopy gas chromatography-mass spectrometry methods. heating temperature were evaluated using chromatography flame ionization detection...
Generative AI technologies have rapidly advanced in recent years, finding extensive applications architectural design, including the generation of building volumes, which represent three-dimensional form buildings. However, current studies this area lack scientific dataset classification methods and primarily focus on single generative models, with limited emphasis model integration. To address these challenges, study presents a comprehensive workflow that integrates multiple spatial...
In this paper, we propose a novel Faster R-CNN based method to detect multiscale objects in very high resolution optical remote sensing images. Firstly, pre-trained CNN is used extract features from an input image; and then set of object candidates are generated. To efficiently with various scales, design hierarchical selective filtering (HSF) layer map different scales the same scale space. The HSF can be applied on both region proposal subsequent detection network. More importantly, it...
Detecting dense buildings without elevation information is an important and challenging task in remote sensing applications. In this paper, we present a novel cascaded deep neural network architecture, incorporating multi -stage region proposal detection Hough transform to obtain better mid-level semantic for man-made objects. This proposed can be trained end-to-end by multi-loss jointly. We train test it on large building dataset collected from Google Earth, including urban, suburban rural...
In recent years, deep learning methods have achieved great success for vehicle detection tasks in aerial imagery. However, most existing focus only on extracting latent target features, and rarely consider the scene context as vital prior knowledge. this letter, we propose a attention-based fusion network (SCAF-Net), to fuse of vehicles into an end-to-end network. First, novel strategy, patch cover, keep original information raw images large scale much possible. Next, use improved YOLO-v3...
Object detection algorithms are pivotal components of unmanned aerial vehicle (UAV) imaging systems, extensively employed in complex fields. However, images captured by high-mobility UAVs often suffer from motion blur cases, which significantly impedes the performance advanced object algorithms. To address these challenges, we propose an innovative algorithm specifically designed for blurry images, named DREB-Net (Dual-stream Restoration Embedding Blur-feature Fusion Network). First,...
With the continuous progress of remote sensing image object detection tasks in recent years, researchers this field have gradually shifted focus their research from horizontal to study arbitrary directions. It is worth noting that some properties are different during oriented yet notice much. This article presents design a straightforward and efficient arbitrary-oriented system, leveraging inherent orientation task, including rotation angle box aspect ratio. In low ratio objects, little...
As an important research branch in data mining, outlier detection has been widely used equipment operation monitoring and system control. Power is playing increasingly vital role power systems. Density peak clustering (DPC) a simple efficient density-based algorithm with good application prospect. Nevertheless, the results by DPC can be greatly influenced cutoff distance, indicating that are highly sensitive to this parameter. To address shortcomings of take characteristics into...
Concealed object detection (COD) in cluttered scenes is significant for various image processing applications. However, due to that concealed objects are always similar their background, it extremely hard distinguish them. Here, the major obstacle tiny feature differences between inside and outside boundary region, which makes trouble existing COD methods achieve accurate results. In this paper, considering surrounding environment information can be well utilized identify objects, thus, we...
Gaussian Boson Sampling (GBS) exhibits a unique ability to solve graph problems, such as finding cliques in complex graphs. It is noteworthy that many drug discovery tasks can be viewed the clique-finding process, making them potentially suitable for quantum computation. However, perform these their quantum-enhanced form, large-scale hardware with universal programmability essential, which yet achieved even most advanced GBS devices. Here, we construct time-bin encoded photonic processor...
The detection of objects such as vehicle, airplane and ship is a fundamental problem in optical remote-sensing(ORS) image process. Despite great success has achieved by migrating nature methods to the remote sensing field, some challenges hardware limit environments still remain be solved, e.g., space-borne UAV-borne hardware. We proposed low-computational network digging several prior knowledge field. By focusing on certain ground sample distance(gsd) single target class, method gains high...
Differential privacy is a new protection technology, which defines strict and strong model, by adding noise data distortion to achieve the purpose of protection.Frequent pattern mining an important field in mining, its find frequent patterns set, but content model itself, rules, counting information likely lead leaking sensitive information.This paper presents item sets method based on differential privacy, named DPFM, adopts strategy combined with Laplace system index system, realizing...
Aimed at the problem that traditional methods fail to deal with malicious attacks under arbitrary background knowledge during process of massive data analysis, an improved Apriori algorithm preserving differential privacy, combining Laplace mechanism mine pattern sensitive information in framework Spark is proposed.Furthermore, it's theoretically proved meet ε-differential privacy spark.Finally, experimental results show guaranteeing availability, our proposed has advantages over protection...