A5016469515- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Multimodal Machine Learning Applications
- Human Pose and Action Recognition
- Advanced Neural Network Applications
- Image Retrieval and Classification Techniques
- Domain Adaptation and Few-Shot Learning
- Topic Modeling
- Gait Recognition and Analysis
- Face recognition and analysis
- Video Analysis and Summarization
- Visual Attention and Saliency Detection
- Remote-Sensing Image Classification
- Infrared Target Detection Methodologies
- Text and Document Classification Technologies
- Advanced Measurement and Detection Methods
- Natural Language Processing Techniques
- Fire Detection and Safety Systems
- Advanced Image Fusion Techniques
- Sentiment Analysis and Opinion Mining
- Image Enhancement Techniques
- Advanced Graph Neural Networks
- Industrial Vision Systems and Defect Detection
- Adversarial Robustness in Machine Learning
- Advanced Text Analysis Techniques
Guangxi Normal University
2016-2025
Guilin University of Electronic Technology
2018
Guangxi University of Science and Technology
2017
Shanghai Jiao Tong University
2012-2013
Most deep trackers still follow the guidance of siamese paradigms and use a template that contains only target without any contextual information, which makes it difficult for tracker to cope with large appearance changes, rapid movement, attraction from similar objects. To alleviate above problem, we propose long-term context attention (LCA) module can perform extensive information fusion on its frames, calculate correlation while enhancing features. The complete location as well state...
ABSTRACT In order to solve the problem that existing PM 2.5 concentration prediction methods ignore spatial and temporal influencing factors of concentration, this paper constructs a characteristic factor based on maximum information coefficient, proposes CNN‐LSTM combined model multi‐feature fusion, which transforms abstract into quantifiable features. The has good feature extraction ability strong capture short‐term transient long‐range dependent in time series data, improves performance...
A good document summary should summarize the core content of text. Research on automatic text summarization attempts to solve this problem. The encoder-decoder model is widely used in research. Soft attention obtain required contextual semantic information during decoding. However, due lack access key features, generated deviates from content. In paper, we proposed an based a double pointer network (DAPT). DAPT, self-attention mechanism collects encoder, soft and generate more coherent...
Most existing image captioning methods use only the visual information of to guide generation captions, lack guidance effective scene semantic information, and current attention mechanism cannot adjust focus intensity on image. In this article, we first propose an improved model. At each timestep, calculated coefficient through context model, then automatically adjusted extract more accurate information. addition, represented knowledge topic words related scene, added them language We used...
The traditional canny edge detection algorithm has its limitations in the aspect of antinoise interference, and it is susceptible to factors such as light. To solve these defects, Canny based on morphological improvement was proposed applied agricultural products. First, uses open close operation morphology form a filter instead Gaussian filter, which can remove image noise strengthen protection edge. Second, operator improved increase horizontal vertical templates 45° 135° improve...
<title>Abstract</title> In intelligent transportation systems, the accurate and real-time detection recognition of traffic signs are crucial for autonomous assisted driving. Despite improvements in efficiency accuracy existing deep learning object algorithms, challenges remain detecting small objects, handling multi-scale targets, achieving low computational resource environments. To address these challenges, we propose a lightweight YOLOv10n-based method that incorporates Hybrid Attention...