A5056659804- Functional Brain Connectivity Studies
- Video Surveillance and Tracking Methods
- Neural dynamics and brain function
- Complex Network Analysis Techniques
- Advanced Image and Video Retrieval Techniques
- Advanced Neuroimaging Techniques and Applications
- Advanced Algorithms and Applications
- Face and Expression Recognition
- Advanced Graph Neural Networks
- GNSS positioning and interference
- Gait Recognition and Analysis
- Robotics and Sensor-Based Localization
- Infrared Target Detection Methodologies
- Advanced Clustering Algorithms Research
- EEG and Brain-Computer Interfaces
- Advanced Computational Techniques and Applications
- Remote-Sensing Image Classification
- Advanced Computing and Algorithms
- Image Retrieval and Classification Techniques
- Neural Networks and Applications
- Indoor and Outdoor Localization Technologies
- Wireless Communication Networks Research
- Image Processing Techniques and Applications
- Neuroscience and Neural Engineering
- Recommender Systems and Techniques
Jiangsu University
2014-2025
Tianjin Medical University
2025
State Key Laboratory of Nuclear Physics and Technology
2024
Peking University
2024
East China Normal University
2023
Shenzhen Polytechnic
2023
Air Force Engineering University
2009-2022
Fudan University
2012-2017
Northwestern Polytechnical University
2008-2009
UtopiaCompression (United States)
2009
Facial Expression Recognition (FER) in the wild poses significant challenges due to realistic occlusions, illumination, scale, and head pose variations of facial images. In this article, we propose an Edge-AI-driven framework for FER. On algorithms aspect, two attention modules, Arbitrary-oriented Spatial Pooling (ASP) Scalable Frequency (SFP), effective feature extraction improve classification accuracy. systems edge-cloud joint inference architecture FER achieve low-latency inference,...
Video-based person re-identification aims to associate the video clips of same across multiple non-overlapping cameras. Spatial-temporal representations can provide richer and complementary information between frames, which are crucial distinguish target when occlusion occurs. This paper proposes a novel Pyramid Spatial-Temporal Aggregation (PSTA) framework aggregate frame-level features progressively fuse hierarchical temporal into final video-level representation. Thus, short-term...
In recent years, multiview clustering algorithms have achieved promising performance by exploiting the complementarity and consistency of different views. However, many spectral methods only focus on consistent information views, time cost feature decomposition is expensive. Moreover, these also require postprocessing (e.g., k-means) to obtain final results. To overcome limitations simultaneously, we propose a novel algorithm. First, method removes inconsistent views through cross-view...
Driven by the potential application to wireless communications, intensive research efforts have been made on study of various selective combining (SC) schemes in past decade. Nevertheless, regardless its practical importance, performance analysis multi-branch SC over spatially correlated fading channels is not available literature except for simplest case dual diversity. The major difficulty lies fact that has root theory order statistics, and yet systematic methodology developed mainly...
Multi-spectral object Re-identification (ReID) aims to retrieve specific objects by leveraging complementary information from different image spectra. It delivers great advantages over traditional single-spectral ReID in complex visual environment. However, the significant distribution gap among spectra poses challenges for effective multi-spectral feature representations. In addition, most of current Transformer-based methods only utilize global class tokens achieve holistic retrieval,...
Planar object tracking is an actively studied problem in vision-based robotic applications. While several benchmarks have been constructed for evaluating state-of-the-art algorithms, there a lack of video sequences captured the wild rather than constrained laboratory environment. In this paper, we present carefully designed planar benchmark containing 210 videos 30 objects sampled natural particular, each object, shoot seven involving various challenging factors, namely scale change,...
Growing evidence indicates that autism spectrum disorder (ASD) is a neuropsychological disconnection syndrome can be analyzed using various complex network metrics used as pathology biomarkers. Recently, community detection and analysis rooted in the graph theories have been introduced to investigate changes resting-state functional structure under neurological pathologies. However, potential of hidden patterns modular organization networks derived from magnetic resonance imaging predict...
In recent years, graph-based deep learning algorithms have attracted widespread attention in the field of consumer electronics. Still, most current graph neural networks are based on supervised or semi-supervised learning, which often relies true labels given samples as auxiliary information. To solve this problem, we propose a Deep Self-Supervised Attention Convolution Autoencoder Graph Clustering (DSAGC) model and use it for social clustering. We divide proposed into two parts: pretext...
Visible-infrared person re-identification (VI-ReID) task is to retrieve the same pedestrian across visible and infrared modalities. The existing transformer-based works are constrained by inherent structure of ViT that feature collapse in deeper layers over-globalization extracted features, resulting incomplete learning local low-level features. However, these features instrumental representing identifying elements within visible-infrared images more comprehensively, which increases accuracy...
It is unsatisfied to diagnose brain disorders based on subjective judgment. In this paper, we proposed a novel method classify autism and normal subjects objectively automatically. The firstly detects community structure in network of every subject. NMI statistic matrix, which can effectively represent the features all subject certain dataset, was developed then imported into denoising autoencoder classify. We tested our three datasets. results show that accuracy higher than traditional one....
Abstract Deep graph clustering is an unsupervised learning task that divides nodes in a into disjoint regions with the help of auto-encoders. Currently, such methods have several problems, as follows. (1) The deep method does not effectively utilize generated pseudo-labels, resulting sub-optimal model training results. (2) Each cluster has different confidence level, which affects reliability pseudo-label. To address these we propose Self-supervised Attribute Graph Clustering (DSAGC) to...