A5062168363- Computer Graphics and Visualization Techniques
- Advanced Vision and Imaging
- Domain Adaptation and Few-Shot Learning
- Robotic Path Planning Algorithms
- Advanced Graph Neural Networks
- Multimodal Machine Learning Applications
- Robotics and Sensor-Based Localization
- Advanced Neural Network Applications
- Topic Modeling
- Advanced Image Processing Techniques
- Advanced Image and Video Retrieval Techniques
- Simulation and Modeling Applications
- Generative Adversarial Networks and Image Synthesis
- Image and Signal Denoising Methods
- Video Surveillance and Tracking Methods
- Image Retrieval and Classification Techniques
- Metaheuristic Optimization Algorithms Research
- Remote-Sensing Image Classification
- Text and Document Classification Technologies
- 3D Shape Modeling and Analysis
- Mobile Ad Hoc Networks
- Distributed and Parallel Computing Systems
- Satellite Communication Systems
- Image Enhancement Techniques
- Remote Sensing and Land Use
Institute of Software
2016-2025
Chinese Academy of Sciences
2015-2025
University of Chinese Academy of Sciences
2008-2025
Institute of Biophysics
2024
Yangzhou University
2024
Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou)
2021-2022
Institute of Oceanology
2007-2021
Shandong Institute of Automation
2021
South China Sea Institute Of Oceanology
2021
University Health Network
2019
Based on evolutionary computation, a novel real-time route planner for unmanned air vehicles is presented. In the planner, individual candidates are evaluated with respect to workspace so that computation of configuration space not required. The incorporates domain-specific knowledge, can handle unforeseeable changes environment, and take into account different kinds mission constraints such as minimum leg length flying altitude, maximum turning angle, fixed approach vector goal position....
Multiple sclerosis (MS) is a chronic and debilitating autoimmune disease, characterized by inflammatory demyelination in the nervous tissue subsequent neurological dysfunction. Spermidine, natural polyamine, has been shown to affect inflammation some experimental models. We show here that spermidine could alleviate encephalomyelitis (EAE), model for MS, through regulating infiltration of CD4+ T cells macrophages central system. Unexpectedly, we found treatment MOG-specific did not their...
Recently, significant progress has been made in masked image modeling to catch up language modeling. However, unlike words NLP, the lack of semantic decomposition images still makes autoencoding (MAE) different between vision and language. In this paper, we explore a potential visual analogue words, i.e., parts, integrate information into training process MAE by proposing Semantic-Guided Masking strategy. Compared widely adopted random masking, our masking strategy can gradually guide...
In this work, we introduce a novel perspective, i.e., dimensional analysis, to address the challenge of communication efficiency in Multi-Agent Reinforcement Learning (MARL). Our findings reveal that simply optimizing content and timing at sending end is insufficient fully resolve issues. Even after applying optimized gated messages, redundancy confounders still persist integrated message embeddings receiving end, which negatively impact quality decision-making. To these challenges, propose...
Few-shot learning models learn representations with limited human annotations, and such a paradigm demonstrates practicability in various tasks, e.g., image classification, object detection, etc. However, few-shot detection methods suffer from an intrinsic defect that the training data makes model cannot sufficiently explore semantic information. To tackle this, we introduce knowledge distillation to paradigm. We further run motivating experiment, which process of distillation, empirical...
Making the transition to a new architecture is never easy. Users want keep running their favorite applications as they normally would, without stopping adapt them different platform. For some legacy problem more severe. Without all source code, it well-nigh impossible recompile application Binary translation helps this process because automatically converts binary code from one instruction set another need for high-level code. However, choices force trade-offs between form of interpretation...
Graph representation learning methods are highly effective in handling complex non-Euclidean data by capturing intricate relationships and features within graph structures. However, traditional face challenges when dealing with heterogeneous graphs that contain various types of nodes edges due to the diverse sources nature data. Existing neural networks (HGNNs) have shown promising results but require prior knowledge node edge unified feature formats, which limits their applicability. Recent...
Although vision Transformers have achieved excellent performance as backbone models in many tasks, most of them intend to capture global relations all tokens an image or a window, which disrupts the inherent spatial and local correlations between patches 2D structure. In this paper, we introduce simple Transformer named SimViT, incorporate structure information into Transformers. Specifically, Multi-head Central Self-Attention(MCSA) instead conventional Self-Attention highly relations. The...
The prevailing graph neural network models have achieved significant progress in representation learning. However, this paper, we uncover an ever-overlooked phenomenon: the pre-trained learning model tested with full graphs underperforms well-pruned graphs. This observation reveals that there exist confounders graphs, which may interfere semantic information, and current methods not eliminated their influence. To tackle issue, propose Robust Causal Graph Representation Learning (RCGRL) to...
Micro-expressions (MEs) are involuntary movements revealing people's hidden feelings, which has attracted numerous interests for its objectivity in emotion detection. However, despite wide applications various scenarios, micro-expression recognition (MER) remains a challenging problem real life due to three reasons, including (i) data-level: lack of data and imbalanced classes, (ii) feature-level: subtle, rapid changing, complex features MEs, (iii) decision-making-level: impact individual...
Deep neural networks, such as Faster R-CNN, have been widely used in object detection.However, deep networks usually require a large-scale dataset to achieve desirable performance.For the specific application, UAV detection, training data is extremely limited practice.Since annotating plenty of images manually can be very resource intensive and time consuming, instead, we use PBRT render large number photorealistic high variation within reasonable time.Using ensures realism rendered images,...