A5100424729- Advanced Image and Video Retrieval Techniques
- Anomaly Detection Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Video Surveillance and Tracking Methods
- Neural Networks Stability and Synchronization
- Adversarial Robustness in Machine Learning
- Stability and Control of Uncertain Systems
- Digital Media and Visual Art
- Speech Recognition and Synthesis
- Text and Document Classification Technologies
- Face and Expression Recognition
- Hepatitis C virus research
- Advanced Clustering Algorithms Research
- Advanced Measurement and Detection Methods
- Infrared Target Detection Methodologies
- Speech and Audio Processing
- Epilepsy research and treatment
- Image and Signal Denoising Methods
- Business Process Modeling and Analysis
- Machine Learning and ELM
- Advanced Decision-Making Techniques
- Advanced Computational Techniques and Applications
- Stability and Controllability of Differential Equations
- Fault Detection and Control Systems
- Educational Technology and Assessment
NetEase (China)
2024-2025
China University of Geosciences
2014-2024
Sichuan University
2022-2024
China Geological Survey
2024
China Institute of Water Resources and Hydropower Research
2022-2023
University of Science and Technology of China
2023
Huawei Technologies (China)
2022
BOE Technology Group (China)
2022
China Automotive Technology and Research Center
2021-2022
Beijing Institute of Graphic Communication
2016
In recent, some robust contrastive multi-view clustering (MvC) methods have been proposed, which construct data pairs from neighborhoods to alleviate the false negative issue, i.e., intra-cluster samples are wrongly treated as pairs. Although promising performance has achieved by these methods, issue is still far addressed and positive emerges because all in- out-of-neighborhood simply negative, respectively. To address issues, we propose a novel method, dubbed decoupled with high-order...
Interictal epileptiform discharge (IED) and its spatial distribution are critical for the diagnosis, classification, treatment of epilepsy. Existing publicly available datasets suffer from limitations such as insufficient data amount lack information. In this paper, we present a comprehensive EEG dataset containing annotated interictal epileptic 84 patients, each contributing 20 minutes continuous raw recordings, totaling 28 hours. IEDs states consciousness (wake/sleep) were meticulously by...
In this paper, the exponential stability in p th( > 1 )-moment for neutral stochastic Markov systems with time-varying delay is studied. The derived conditions comprise two forms: 1) delay-independent criteria which are obtained by establishing an integral inequality and 2) delay-dependent captured using theory of functional differential equations. As its applications, results used to investigate neural networks switching, globally adaptive synchronization complex dynamical respectively. On...
Multi-Target Multi-Camera tracking is a fundamental task for intelligent traffic systems. The track 1 of AI City Challenge 2022 aims at the city-scale multi-camera vehicle task. In this paper we propose an accurate system composed 4 parts, including: (1) State-of-the-art detection and re-identification models feature extraction. (2) Single camera tracking, where introduce augmented tracks prediction multi-level association method on top tracking-by-detection paradigm.(3) Zone-based...
Anomaly detection on multivariate time series (MTS) is an important research topic in data mining, which has a wide range of applications information technology, financial management, manufacturing system, and so on. However, the state-of-the-art unsupervised deep learning models for MTS anomaly are vulnerable to noise have poor performance training containing anomalies. In this article, we propose novel Self-Training based Detection with Generative Adversarial Network (GAN) model called...
In this paper, we study a novel and widely existing problem in graph matching (GM), namely, Bi-level Noisy Correspondence (BNC), which refers to node-level noisy correspondence (NNC) edge-level (ENC). brief, on the one hand, due poor recognizability viewpoint differences between images, it is inevitable inaccurately annotate some keypoints with offset confusion, leading mismatch two associated nodes, i.e., NNC. On other node-to-node will further contaminate edge-to-edge correspondence, thus...
Recently, with the advent of Metaverse and AI Generated Content, cross-modal retrieval becomes popular a burst 2D 3D data. However, this problem is challenging given heterogeneous structure semantic discrepancies. Moreover, imperfect annotations are ubiquitous ambiguous content, thus inevitably producing noisy labels to degrade learning performance. To tackle problem, paper proposes robust 2D-3D framework (RONO) robustly learn from multimodal Specifically, one novel Robust Discriminative...
Multi-scale architectures have shown effectiveness in a variety of tasks thanks to appealing cross-scale complementarity. However, existing treat different scale features equally without considering the scale-specific characteristics, \textit{i.e.}, within-scale characteristics are ignored architecture design. In this paper, we reveal missing piece for multi-scale design and accordingly propose novel Multi-Scale Adaptive Network (MSANet) single image denoising. Specifically, MSANet...
Cross-domain image retrieval aims at retrieving images across different domains to excavate cross-domain classificatory or correspondence relationships. This paper studies a less-touched problem of retrieval, i.e., unsupervised considering the following practical assumptions: (i) no relationship, and (ii) category annotations. It is challenging align bridge distinct without correspondence. To tackle challenge, we present novel Correspondence-free Domain Alignment (CoDA) method effectively...
Abstract Multi-objective optimization algorithms have shown effectiveness on problems with two or three objectives. As the number of objectives increases, proportion non-dominated solutions increases rapidly, resulting in insufficient selection pressure. Nevertheless, pressure usually leads to loss convergence, too intense often results a lack diversity. Hence, balancing convergence and diversity remains challenging problem many-objective problems. To remedy this issue, evolutionary...
To enhance deep learning-based automated interictal epileptiform discharge (IED) detection, this study proposes a multimodal method, vEpiNet, that leverages video and electroencephalogram (EEG) data. Datasets comprise 24 931 IED (from 484 patients) 166 094 non-IED 4-second video-EEG segments. The data is processed by the proposed patient detection with frame difference Simple Keypoints (SKPS) capturing patients' movements. EEG EfficientNetV2. features are fused via multilayer perceptron. We...
Recently, image-text matching has attracted more and attention from academia industry, which is fundamental to understanding the latent correspondence across visual textual modalities. However, most existing methods implicitly assume training pairs are well-aligned while ignoring ubiquitous annotation noise, a.k.a noisy (NC), thereby inevitably leading a performance drop. Although some attempt address such they still face two challenging problems: excessive memorizing/overfitting unreliable...
In this study, by constructing an appropriate Lyapunov–Krasovskii functional, one delay‐dependent exponential stability criterion of neutral stochastic system with multiple time‐varying delays is obtained in terms linear matrix inequalities. The significant contributions study are that the fewer variables results involved and some less computational burdens imposed. particular, can greatly reduce conservatism when time‐invariant case considered. Finally, three illustrative numerical examples...
In this paper, we study a novel and widely existing problem in graph matching (GM), namely, Bi-level Noisy Correspondence (BNC), which refers to node-level noisy correspondence (NNC) edge-level (ENC). brief, on the one hand, due poor recognizability viewpoint differences between images, it is inevitable inaccurately annotate some keypoints with offset confusion, leading mismatch two associated nodes, i.e., NNC. On other node-to-node will further contaminate edge-to-edge correspondence, thus...