A5100444159- Video Surveillance and Tracking Methods
- Human Pose and Action Recognition
- Blockchain Technology Applications and Security
- Cloud Computing and Resource Management
- IoT and Edge/Fog Computing
- Cryptography and Data Security
- Retinal Imaging and Analysis
- Cloud Data Security Solutions
- Advanced Memory and Neural Computing
- Adversarial Robustness in Machine Learning
- Recommender Systems and Techniques
- Retinal and Optic Conditions
- Advanced Graph Neural Networks
- Software-Defined Networks and 5G
- Computational Drug Discovery Methods
- Machine Learning in Materials Science
- Anomaly Detection Techniques and Applications
- Caching and Content Delivery
- Advanced Neural Network Applications
- Fire Detection and Safety Systems
- Human Mobility and Location-Based Analysis
- Automated Road and Building Extraction
- Security and Verification in Computing
- FinTech, Crowdfunding, Digital Finance
- Advanced Malware Detection Techniques
Beihang University
2020-2024
State Key Laboratory of Software Development Environment
2023
Shandong University of Science and Technology
2023
Beijing Advanced Sciences and Innovation Center
2020
Recently, most multiple object tracking (MOT) algorithms adopt the idea of tracking-by-detection. Relevant research shows that performance detector obviously affects tracker, while improvement is gradually slowing down in recent years. Therefore, trackers using tracklet (short trajectory) are proposed to generate more complete trajectories. Although there various generation algorithms, fragmentation problem still often occurs crowded scenes. In this paper, we introduce an iterative...
Multiobject tracking is a basic task in video analysis. Due to the strict requirements on efficiency and resource consumption, most of applications edge devices are online or near-online methods. Besides motion modeling, appearance information also widely used for tracking. However, influence occlusion usually ignored. In this article, spatial-temporal co-occurrence constraints (STCCs) features introduced resist occlusions by exploring rich spatial temporal tracklets. addition, novel...
Person reidentification (ReID) is an important application of Internet Things (IoT). ReID recognizes pedestrians across camera views at different locations and time, which usually treated as a ranking task. An essential part this task the hard sample mining. Technically, two strategies could be employed, i.e., global mining local For former, samples are mined within entire training set, while for latter, it done in mini-batches. In literature, most existing methods operate locally. Examples...
Designing new molecules is essential for drug discovery and material science. Recently, deep generative models that aim to model molecule distribution have made promising progress in narrowing down the chemical research space generating high-fidelity molecules. However, current only focus on modeling 2-D bonding graphs or 3-D geometries, which are two complementary descriptors The lack of ability jointly them limits improvement generation quality further downstream applications. In this...
Data association is one of the key research in tracking-by-detection framework. Due to frequent interactions among targets, there are various relationships trajectories crowded scenes which leads problems data association, such as ambiguity, omission, etc. To handle these problems, we propose hypothesis-testing based tracking (HTBT) framework build potential associations between target by constructing and testing hypotheses. In addition, a spatio-temporal interaction graph (STIG) model...
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Overprivilege Attack</i> , a widely reported phenomenon in IoT that accesses unauthorized or excessive resources, is notoriously hard to prevent, trace and mitigate. In this paper, we propose TBAC, Tokoin-Based Access Control model enabled by blockchain Trusted Execution Environment (TEE) technologies, offer fine-grained access control strong auditability for IoT. TBAC materializes the virtual...
Vehicle reidentification, aiming at identifying vehicles across images, has drawn a lot of attention and made significant achievements in recent years. However, vehicle reidentification remains challenging task caused by severe appearance changes due to different orientations. In practice, the result is greatly influenced pose vehicles, we call this influence as barrier problem. One way address problem train feature representation that invariant for various poses. To end, present robust...
Zero-Knowledge Succinct Non-Interactive Argument of Knowledge (zk-SNARK) is a practical zero-knowledge proof system for Rank-1 Constraint Satisfaction (R1CS), enabling privacy preservation and addressing the previous scalability concerns on proofs. Existing constructions zk-SNARKs require huge memory overhead to generate proofs in that size zk-SNARK circuit can be large even very simple use case, which limits applications regular resource-constrained users. To reduce utilization zk-SNARKs,...
Automatic clinical diagnosis of retinal diseases has emerged as a promising approach to facilitate discovery in areas with limited access specialists. We propose novel visual-assisted hybrid model based on the support vector machine (SVM) and deep neural networks (DNNs). The incorporates complementary strengths DNNs SVM. Furthermore, we present new retina label collection for ophthalmology incorporating 32 classes. Using EyeNet, our achieves 89.73% accuracy performance is comparable...
We consider a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$K$ </tex-math></inline-formula> -armed bandit problem in general graphs where agents are arbitrarily connected and each of them has limited memorizing capabilities communication bandwidth. The goal is to let the eventually learn best arm. Although recent studies show power collaboration among improving efficacy learning, it assumed these that...
The proliferation of edge computing brings new challenges due to the complexity decentralized networks. Software-defined networking (SDN) takes advantage pro-grammability and flexibility in handling complicated However, it remains a problem designing both trusted scalable SDN control plane, which is core component architecture for computing. In this paper, we propose Curb, novel group-based plane that seamlessly integrates blockchain BFT consensus ensure byzantine fault tolerance,...
With the continuous improvement of information infrastructures, academia and industry have been constantly exploring new computing paradigms to fully exploit powers. In this paper, we propose Meta Computing, a paradigm that aims utilize all available resources hooked on Internet, provide efficient, fault-tolerant, personalized services with strong security privacy guarantee, virtualize Internet as giant computer, is, "Network-as-a-Computer, NaaC", or "Meta Computer" for short, any task...
With the continuous improvement of information infrastructures, academia and industry have been constantly exploring new computing paradigms to fully exploit powers. In this paper, we propose Meta Computing, a paradigm that aims utilize all available resources hooked on Internet, provide efficient, fault-tolerant, personalized services with strong security privacy guarantee, virtualize Internet as giant computer, is, ``Network-as-a-Computer, NaaC'', or ``Meta Computer'' for short, any task...