A5101721258- Advanced Neural Network Applications
- Remote Sensing and LiDAR Applications
- Microbial Community Ecology and Physiology
- 3D Shape Modeling and Analysis
- Educational Technology and Assessment
- Autonomous Vehicle Technology and Safety
- Robotic Path Planning Algorithms
- Robotics and Sensor-Based Localization
- Genomics and Phylogenetic Studies
- Handwritten Text Recognition Techniques
- Cancer Mechanisms and Therapy
- Sport and Mega-Event Impacts
- Remote Sensing and Land Use
- Porphyrin Metabolism and Disorders
- Polar Research and Ecology
- Optimization and Search Problems
- Wastewater Treatment and Nitrogen Removal
- 3D Surveying and Cultural Heritage
- Clay minerals and soil interactions
- Wireless Sensor Networks and IoT
- Protist diversity and phylogeny
- Protein Tyrosine Phosphatases
- Simulation and Modeling Applications
- Biomimetic flight and propulsion mechanisms
- Intelligent Tutoring Systems and Adaptive Learning
State Key Laboratory of Satellite Ocean Environment Dynamics
2022-2024
Hainan University
2022-2024
Ministry of Natural Resources
2022-2024
Second Institute of Oceanography
2022-2024
Xiamen University
2020-2024
RELX Group (Netherlands)
2024
Nanjing University of Aeronautics and Astronautics
2022-2023
Wenzhou Central Hospital
2023
China University of Mining and Technology
2019
Internationale Akademie Berlin
1968
In this study, AOA communities along a river in southern China were characterized, and metagenome-assembled genomes (MAGs) of novel clade also obtained. Based on the characterization genomes, study suggests adaptation AOAs to estuarine environments, providing new information ecology nitrogen cycle contaminated environments.
In recent years, there has been a rapid rise in interest trajectory prediction the field of autonomous driving. However, domain generalization current works is often neglected, and their performance tends to degrade when transferred different dataset or scenario. this article, we present solution problem that accounts for realistic conditions necessary Specifically, identify velocity environment as possible causes decline generalization. Then, propose incorporating module refinement issue....
Phospholipid phosphatase related 4 (PLPPR4), a neuron-specific membrane protein located at the postsynaptic density of glutamatergic synapses, is putative regulator neuronal plasticity. However, PLPPR4 dysfunction has not been linked to genetic disorders. In this study, we report three unrelated patients with intellectual disability (ID) or autism spectrum disorder (ASD) who harbour de novo heterozygous copy number loss in 1p21.2p21.3, nonsense mutation (NM_014839, c.4C > T, p.Gln2*) and...
Semantic segmentation plays a crucial role in enabling intelligent vehicles to perceive and understand their surroundings. However, datasets used for semantic often suffer from data imbalance, where the number of pixels belonging different classes varies significantly. To address this challenge, various novel loss functions have been proposed at class or pixel level counterbalance imbalance. In study, we propose approach mitigate problem subclass perspective. Specifically, identify...
LiDAR point cloud segmentation is one of the most fundamental tasks for autonomous driving scene understanding. However, it difficult existing models to achieve both high inference speed and accuracy simultaneously. For example, voxel-based methods perform well in accuracy, while Bird's-Eye-View (BEV)-based can real-time inference. To overcome this issue, we develop an effective 3D-to-BEV knowledge distillation method that transfers rich from 3D BEV-based models. Our framework mainly...
Abstract Land–ocean interactions greatly impact the evolution of coastal life on earth. However, ancient geological forces and genetic mechanisms that shaped evolutionary adaptations allowed microorganisms to inhabit brackish waters remain largely unexplored. In this study, we infer trajectory ubiquitous heterotrophic archaea Poseidoniales (Marine Group II archaea) presently occurring across global aquatic habitats. Our results show their subgroups had a single origination, dated over 600...
Aiming at the problem that global path optimization cannot be guaranteed in dynamic planning of rotorcraft formation, a static and obstacle avoidance algorithm combining improved whale artificial potential field method is proposed. The results are greatly affected by distribution initial solutions. paper proposes to combine hierarchical system grey wolf with standard algorithm, incorporate first three historical optimal solutions into calculation range improve ability population escape from...
Abstract In order to solve the problems of large number, small size and low detection accuracy vehicle targets in aerial photography, a target algorithm based on improved YOLOv3 is proposed this paper. Firstly, aiming at problem information loss road vehicles, new added.Secondly, better detect targets, 104x104 scale layer added basis three layers traditional yolov3 network structure.The k-means + used cluster data set, ground candidate frame obtained. The function by using Focal...
In this paper, we introduce a novel knowledge distillation approach for the semantic segmentation task. Unlike previous methods that rely on power-trained teachers or other modalities to provide additional knowledge, our does not require complex teacher models information from extra sensors. Specifically, model training, propose noise label and then incorporate it into input effectively boost lightweight performance. To ensure robustness of against introduced noise, dual-path consistency...
Abstract Land-ocean interactions greatly impacted the evolution of coastal life on Earth. However, geological forces and genetic mechanisms that shaped evolutionary adaptations allowed microorganisms to inhabit brackish waters remain largely unexplored. Here, we infer trajectory ubiquitous heterotrophic archaea Poseidoniales (Marine Group II archaea) across global aquatic habitats. Our results show their subgroups have evolved through rearrangement magnesium transport gene corA conferred...
LiDAR point cloud segmentation is one of the most fundamental tasks for autonomous driving scene understanding. However, it difficult existing models to achieve both high inference speed and accuracy simultaneously. For example, voxel-based methods perform well in accuracy, while Bird's-Eye-View (BEV)-based can real-time inference. To overcome this issue, we develop an effective 3D-to-BEV knowledge distillation method that transfers rich from 3D BEV-based models. Our framework mainly...
Semantic Scene Completion (SSC) aims to jointly generate space occupancies and semantic labels for complex 3D scenes. Most existing SSC models focus on volumetric representations, which are memory-inefficient large outdoor spaces. Point clouds provide a lightweight alternative but benchmarks lack point cloud scenes with labels. To address this, we introduce PointSSC, the first cooperative vehicle-infrastructure benchmark scene completion. These exhibit long-range perception minimal...