A5072713174- Advanced Image Processing Techniques
- Reinforcement Learning in Robotics
- Advanced Neural Network Applications
- Elevator Systems and Control
- Advanced Image Fusion Techniques
- Industrial Gas Emission Control
- Image and Signal Denoising Methods
- Odor and Emission Control Technologies
- Advanced Image and Video Retrieval Techniques
- Supply Chain and Inventory Management
- Wastewater Treatment and Nitrogen Removal
- Advanced Mathematical Physics Problems
- Infrastructure Maintenance and Monitoring
- Image Processing Techniques and Applications
- Stability and Controllability of Differential Equations
- Industrial Vision Systems and Defect Detection
- Digital Media and Visual Art
- Remote-Sensing Image Classification
- Evolutionary Algorithms and Applications
- Advanced Vision and Imaging
- Domain Adaptation and Few-Shot Learning
- Iterative Learning Control Systems
- Image Retrieval and Classification Techniques
- Simulation Techniques and Applications
- Robotics and Sensor-Based Localization
Linyi University
2008-2024
University Town of Shenzhen
2024
Tsinghua University
2024
Tsinghua–Berkeley Shenzhen Institute
2024
Real-time remote sensing segmentation technology is crucial for unmanned aerial vehicles (UAVs) in battlefield surveillance, land characterization observation, earthquake disaster assessment, etc., and can significantly enhance the application value of UAVs military civilian fields. To realize this potential, it essential to develop real-time semantic methods that be applied resource-limited platforms, such as edge devices. The majority mainstream rely on convolutional neural networks (CNNs)...
Defect detection plays a crucial role in ensuring the surface quality of steel, as it impacts both downstream production and overall final product. However, recognizing defects steel has always been challenging due to small occurrence area, diverse types deformations, various defects. In this article, we propose novel lightweight defect model named YOLOv5s-DNF, which achieves high performance while maintaining architecture. Specifically, enhance C3 structure YOLOv5 by incorporating...
In the present paper, we show that global solution to (partially) damped Klein-Gordon equation on three dimensional Euclidean space with small data decays exponentially. The key ingredients in proof are: Morawetz-type estimates for solutions and Ruiz's unique continuation principle wave equations.
High-resolution Satellite cloud images play a crucial role in weather analysis and forecast. Improving the resolution of with super-resolution (SR) methods facilitates system to identify locate geographical information. In this paper, several SR have been verified on natural color (NCCI) dataset. We propose multi-scale interaction convolution (MIC), which realizes adaptive aggregation features under different reception fields. The semi-dense residual is designed promote dissemination Based...
Direct Preference Optimization (DPO) has recently expanded its successful application from aligning large language models (LLMs) to text-to-image with human preferences, which generated considerable interest within the community. However, we have observed that these approaches rely solely on minimizing reverse Kullback-Leibler divergence during alignment process between fine-tuned model and reference model, neglecting incorporation of other constraints. In this study, focus extending in...
Since the advent of Transformer, Vision Transformer (ViT) model has become optimal solution for extracting semantic information in remote sensing image segmentation tasks. However, recent research discovered that using larger convolutional kernels to extract global can achieve performance comparable to, or even better than ViT models. This finding inspires us redesign neural network structure and innovatively construct a novel parallel global-local module based on large kernel convolution,...