- Face recognition and analysis
- Emotion and Mood Recognition
- Advanced Graph Neural Networks
- Mental Health via Writing
- Electrostatics and Colloid Interactions
- Anomaly Detection Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Magnetic Properties and Applications
- Thermal Analysis in Power Transmission
- Microfluidic and Bio-sensing Technologies
- Multimodal Machine Learning Applications
- Complex Network Analysis Techniques
- Cerebrovascular and Carotid Artery Diseases
- Power Transformer Diagnostics and Insulation
- Electrical and Bioimpedance Tomography
Hefei University of Technology
2017-2025
It is a common problem that transformers have different degrees of the turn to short‐circuit (TTS). The fault severity depends on location and size short‐circuit. In this paper, position TTS controlled in middle transformer winding. And then, one with 5%, 10% 20% high voltage side phase B was constructed. faults from slight deep are simulated, effect winding analyzed. First, three‐phase modeled by numerical computation method simulated field‐circuit coupling mode. distribution parameters...
Graph Contrastive Learning (GCL) has achieved great success in self-supervised representation learning throughout positive and negative pairs based on graph neural networks (GNNs), where one critical issue lies how to handle the false hard negatives that share large similarity same referenced class as anchor, which is message passing of GNNs exploit structure. However, existing arts either mistakenly identify or miss negatives, hence resulting into poor node representation. Building this,...
Graph-based anomaly detection is currently an important research topic in the field of graph neural networks (GNNs). We find that detection, homophily distribution differences between different classes are significantly greater than those homophilic and heterophilic graphs. For first time, we introduce a new metric called Class Homophily Variance, which quantitatively describes this phenomenon. To mitigate its impact, propose novel GNN model named Edge Generation Graph Neural Network...
Formaldehyde is a toxic water‐soluble organic matter. Its detection methods and limit have been attracting more attentions. An AC electrokinetic (ACEK)‐enhanced capacitive sensing method for rapid of trace formaldehyde in liquid presented using commercial microelectrode chips. By this method, achieved as low 0.01 ppm (μg/ml) within 60 s response time. It proved that the amount common solvents can also be effectively detected. As complement, capacitance change rate dependence on testing...
Depression is a prevalent mental health disorder that significantly impacts individuals' lives and well-being. Early detection intervention are crucial for effective treatment management of depression. Recently, there many end-to-end deep learning methods leveraging the facial expression features automatic depression detection. However, most current overlook temporal dynamics expressions. Although very recent 3DCNN remedy this gap, they introduce more computational cost due to selection...
Motivation: Magnetic Resonance Vessel Wall Imaging is a crucial method for assessing arterial plaques. Goal(s): Achieving rapid and accurate algorithm vascular centerline extraction automatic detection of the pituitary stalk, thereby enabling fast, automatic, quantitative analysis Approach: Proposed point based on V-NET, designed to achieve centerlines localization stalk. Results: Accurate key had been achieved, enhancing precision fast Impact: The accuracy has improved, plaque enhancement...