A5100322958- Visual Attention and Saliency Detection
- Advanced Memory and Neural Computing
- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Adversarial Robustness in Machine Learning
- Image Enhancement Techniques
- Advanced Malware Detection Techniques
- Advanced Computational Techniques and Applications
- Industrial Vision Systems and Defect Detection
- Microplastics and Plastic Pollution
- Cooperative Communication and Network Coding
- Network Security and Intrusion Detection
- Opportunistic and Delay-Tolerant Networks
- Ferroelectric and Negative Capacitance Devices
- CCD and CMOS Imaging Sensors
- Interconnection Networks and Systems
- biodegradable polymer synthesis and properties
- Caching and Content Delivery
- Software-Defined Networks and 5G
- Digital Media and Visual Art
- Technology and Security Systems
- Internet Traffic Analysis and Secure E-voting
- Innovation in Digital Healthcare Systems
- Power Line Inspection Robots
- Network Packet Processing and Optimization
Dalian University of Technology
2016-2025
China University of Geosciences
2024-2025
Wuhan University
2020-2024
PLA Army Engineering University
2022-2024
University of Science and Technology of China
2022-2024
Shanghai Electric (China)
2024
Harbin Institute of Technology
2023
Dalian University
2023
Guizhou University
2023
Hunan Normal University
2023
Glass is very common in our daily life. Existing computer vision systems neglect it and thus may have severe consequences, e.g., a robot crash into glass wall. However, sensing the presence of not straightforward. The key challenge that arbitrary objects/scenes can appear behind glass, content within region typically similar to those it. In this paper, we propose an important problem detecting from single RGB image. To address problem, construct large-scale detection dataset (GDD) design...
As emerging pollutants in the environment, nanoplastics (NPs) can cross biological barriers and be enriched organisms, posing a greatest threat to health of livestock humans. However, size-dependent toxic effects NPs higher mammals remain largely unknown. To determine potential toxicities NPs, we exposed mouse (AML-12) human (L02) liver cell lines vitro, 6-week-old C57BL/6 mice (well-known preclinical model) vivo five different sizes polystyrene (PS-NPs) (20, 50, 100, 200 500 nm). We found...
Machine learning models have achieved state-of-the-art performance in various fields, from image classification to speech recognition. However, such are trained with a large amount of sensitive training data, and typically computationally expensive build. As result, many cloud providers (e.g., Google) launched machine-learning-as-a-service, which helps clients benefit the sophisticated cloud-based machine via accessing public APIs. Such business paradigm significantly expedites simplifies...
To achieve the non-destructive detection of peanut defects, a multi-target identification method based on multispectral system and improved Faster RCNN is proposed in this paper. In terms system, root-mean-square contrast was employed to select characteristic wavelengths for such as mildew spots, mechanical damage, germ peanuts. Then, light source symmetric integrating sphere designed with 2% nonuniformity illumination. improvement, texture-based attention feature enhancement module were...
Laser irradiation, as a kind of post-fabrication method for two-dimensional (2D) materials, is promising way to tune the properties materials and performance corresponding nano-devices. As memristor has been regarded an excellent candidate in-memory devices in next-generation computing system, application laser irradiation developing based on 2D should be explored deeply. Here, tellurene (Te) flakes are exposed 532 nm air atmosphere investigate evolutions surface morphology atom structures...
ABSTRACT Identifying key influential nodes in Earth surface data association networks is crucial for optimizing the use of scientific data. However, challenges such as network size, complexity, and dynamic node influence make this task difficult. While deep learning methods have improved recognition accuracy reduced computational costs complex networks, they still struggle with balancing efficiency accuracy. To address this, we propose DCKH‐CNN, a novel Multimetric Graph‐Based Convolutional...
Abstract Co─Fe layered double hydroxide (LDH) has shown great performance for oxygen evolution reaction (OER), however, the effect of Fe in compounds as well its mechanism promoting OER are still challenge. Herein, bimetallic silicate hydroxides, with a similar structure LDH and more advantages, designed to address issue about performance‐determining site (PDS). In two different coordinated ions cationic layers (Fe 3+ (Oh)) anionic (Td)) distinguished achieve high‐efficient catalysts, where...
Traditionally, research in the field of traffic safety has predominantly focused on two key areas—the identification black spots and analysis accident causation. However, such heavily relies historical records obtained from management department, which often suffer missing or incomplete information. Moreover, these typically offer limited insight into various attributes associated with accidents, thereby posing challenges to comprehensive analyses. Furthermore, collection data incur...
Time-triggered communication protocols, such as time-triggered protocol (TTP) and FlexRay, have the potential to solve many system integration concurrent engineering issues in aerospace industry. This paper investigates scheduling of periodic applications on systems. A novel problem is formulated capture a unique feature commonly existing safety-critical systems, i.e., task graphs running some nodes (i.e., tasks messages) are strictly while others not. To address problem, algorithm called...
Camouflaged objects can be perfectly hidden in the surrounding environment by designing their texture and color. Existing object detection models have high false-negative rates inaccurate localization for camouflaged objects. To resolve this, we improved YOLOv8 algorithm based on feature enhancement. In extraction stage, an edge enhancement module was built to enhance feature. fusion multiple asymmetric convolution branches were introduced obtain larger receptive fields achieve multi-scale...
With the improvement of software security, attacks based on RPC vulnerabilities declined, however, client application have increased. Such includes Web browsers, email and office. The spread malware using these has become a severe threat to todaypsilas Internet. In allusion this kind threat, paper designed an Internet collecting system client-side honeypot. This can not only collect but also detect malicious site. It uses unique network crawler attack techniques source URL, it collects URL...
Cost-volume construction, which accurately computes the similarities between pixels in paired images, is a fundamental kernel of stereo vision processing and has been directly used robotic, autopilot, AR/VR applications. However, large parameter size consecutive data accesses real-time cost-volume construction (>30fps) exerts high demand on memory bandwidth (0.254Tb/s) operation (391GOPs). A promising candidate to resolve bottleneck computation-in-memory (CIM), provides computing parallelism...
Mirrors are everywhere in our daily lives. Existing computer vision systems do not consider mirrors, and hence may get confused by the reflected content inside a mirror, resulting severe performance degradation. However, separating real outside mirror from it is non-trivial. The key challenge that mirrors typically reflect contents similar to their surroundings, making very difficult differentiate two. In this article, we present novel method segment single RGB image. To best of knowledge,...
NDN (Named Data Networking), a promising next-generation architecture, puts named content in the first place of network and is resilient to many existing DDoS attacks. However, Interest Flooding Attack (IFA), typical NDN-specific attack, has been widely recognized as serious threat development NDN. The countermeasures against IFA mainly aim at scenario that attackers send spoofed Interests fairly high rate intermediate routers near can timely detect attack by themselves. Instead, this work...
Depth maps contain geometric clues for assisting Salient Object Detection (SOD). In this paper, we propose a novel Cross-Modal Weighting (CMW) strategy to encourage comprehensive interactions between RGB and depth channels RGB-D SOD. Specifically, three RGB-depth interaction modules, named CMW-L, CMW-M CMW-H, are developed deal with respectively low-, middle- high-level cross-modal information fusion. These modules use Depth-to-RGB Weighing (DW) RGB-to-RGB (RW) allow rich cross-scale among...
Deep Neural Networks (DNNs) are widely used for computer vision tasks. However, it has been shown that deep models vulnerable to adversarial attacks, i.e., their performances drop when imperceptible perturbations made the original inputs, which may further degrade following visual tasks or introduce new problems such as data and privacy security. Hence, metrics evaluating robustness of against attacks desired. previous mainly proposed shallow networks on small-scale datasets. Although Cross...