A5101476848- Advanced Image and Video Retrieval Techniques
- Remote-Sensing Image Classification
- Augmented Reality Applications
- Ammonia Synthesis and Nitrogen Reduction
- Advanced Photocatalysis Techniques
- Image and Video Quality Assessment
- Energy Harvesting in Wireless Networks
- Handwritten Text Recognition Techniques
- Methane Hydrates and Related Phenomena
- Caching and Content Delivery
- nanoparticles nucleation surface interactions
- Visual Attention and Saliency Detection
- Privacy-Preserving Technologies in Data
- Advanced Image Fusion Techniques
- Antenna Design and Analysis
- Remote Sensing and Land Use
- Face recognition and analysis
- Image Processing and 3D Reconstruction
- Advanced Neural Network Applications
- Minerals Flotation and Separation Techniques
- Domain Adaptation and Few-Shot Learning
- E-commerce and Technology Innovations
- Human Motion and Animation
- Antenna Design and Optimization
- Geological Modeling and Analysis
Xi'an University of Architecture and Technology
2025
China Tobacco
2025
Yunnan University
2025
Beijing Institute of Technology
2024
Xidian University
2020-2023
Assessing the geological suitability of urban underground space development is crucial for mitigating risks. Traditional 2D evaluation methods fail to capture complex vertical variations in space, hindering precise planning. This paper presents an innovative 3D-CWC framework, combining a weighted cloud model with three-dimensional modeling, address complexity and uncertainty assessments. The study area, located northern part Kunming’s Second Ring Road, divided into 22 million 25 m × 1 3D...
Geometric diagrams are critical in conveying mathematical and scientific concepts, yet traditional diagram generation methods often manual resource-intensive. While text-to-image has made strides photorealistic imagery, creating accurate geometric remains a challenge due to the need for precise spatial relationships scarcity of geometry-specific datasets. This paper presents MagicGeo, training-free framework generating from textual descriptions. MagicGeo formulates process as coordinate...
Charged solid substrates play a crucial role in influencing the behavior of interfacial nanobubbles, although underlying mechanisms are not yet fully understood. To explore this process greater depth, we employed molecular dynamics (MD) simulations to systematically examine effects charged graphene on morphological evolution, interface structure, and stability thereby revealing intrinsic mechanisms. Our findings indicate that as surface charge density increases, gas-solid interactions...
Recently, the enactment of privacy regulations has promoted rise machine unlearning paradigm. Existing studies mainly focus on sample-wise unlearning, such that a learnt model will not expose user's at sample level. Yet we argue ability selective removal should also be presented attribute level, especially for attributes irrelevant to main task, e.g., whether person recognized in face recognition system wears glasses or age range person. Through comprehensive literature review, it is found...
Autonomous artificial intelligence (AI) agents have emerged as promising protocols for automatically understanding the language-based environment, particularly with exponential development of large language models (LLMs). However, a fine-grained, comprehensive multimodal environments remains under-explored. This work designs an autonomous workflow tailored integrating AI seamlessly into extended reality (XR) applications fine-grained training. We present demonstration training assistant LEGO...
In this paper a new Sparse Tensor Auto-Encoder (STAE) model is proposed to learn latent and discriminative feature representations for saliency detection. By formulating the background patches as holistic high-dimensional tensors learning multi-dimensional dictionary code image patches, coding error can precisely reveal difference between salient object background. Then map be derived by subsequent refinement of representation errors via segmentation. Several benchmark datasets are used...