A5101719575- Advanced Neural Network Applications
- Neural dynamics and brain function
- Neuroscience and Neural Engineering
- Functional Brain Connectivity Studies
- Remote Sensing and LiDAR Applications
- Neurological disorders and treatments
- Advanced Optical Sensing Technologies
- Leaf Properties and Growth Measurement
- Astrophysics and Cosmic Phenomena
- Lightning and Electromagnetic Phenomena
- Precipitation Measurement and Analysis
- Soybean genetics and cultivation
- Brain Tumor Detection and Classification
- Advanced Memory and Neural Computing
- AI in cancer detection
- Parkinson's Disease Mechanisms and Treatments
- 3D Surveying and Cultural Heritage
- Asthma and respiratory diseases
- Atmospheric chemistry and aerosols
- Vehicle License Plate Recognition
- Medical Image Segmentation Techniques
- Odor and Emission Control Technologies
- Smart Agriculture and AI
- Image Processing Techniques and Applications
- Atmospheric and Environmental Gas Dynamics
Zhengzhou University
2014-2025
Peking University
2023
Shandong University
2023
Drexel University
2011-2022
Northeast Agricultural University
2022
In this paper, we propose an improved B-H-Deformable-DETR (Bayesian H-Deformable-DETR) model for the problem of insufficient accuracy and generalization ability in task Few-Shot Object Detection. our study, adopt Bayesian MLP (Multi-Layer Perceptron based on Linear Layers) as prediction layer bounding box regression. By introducing prior distribution uncertainty estimation, transform neural network with deterministic parameters into a probabilistic stochastic properties. This can handle...
Oceanic nitrogen deposition influences marine ecosystem eutrophication and the global carbon cycle. Its future spatiotemporal features still remain unclear driven by changing anthropogenic emissions. Furthermore, existing studies reported air quality climate benefits of ambitious emission reductions, while consequent impacts for ecosystems through atmospheric are unexplored. Here we utilize chemistry transport model GEOS-Chem to evaluate changes in oceanic between 2015 2050 under three CMIP6...
The stem-related phenotype of mature stage soybean is important in material selection. How to improve on traditional manual methods and obtain the more quickly accurately a problem faced by producers. With development smart agriculture, many scientists have explored phenotypes proposed new acquisition methods, but studies are relatively scarce. In this study, we used deep learning method within convolutional neural network detect stem nodes identified structural features through novel...
The Large High Altitude Air Shower Observatory (LHAASO) has three sub-arrays, KM2A, WCDA and WFCTA. flux variations of cosmic ray air showers were studied by analyzing the KM2A data during thunderstorm on 10 June 2021. number shower events that meet trigger conditions increases significantly in atmospheric electric fields, with maximum fractional increase 20%. rates (increases or decreases) are found to be strongly dependent primary zenith angle. secondary particles significantly, following...
Ammonia (NH3) emission control has been advocated for its potential to mitigate PM2.5 air pollution, yet quantifications at city levels are limited. Here we develop high-resolution (3 km) bottom-up inventories of agricultural NH3 in the Beijing-Tianjin-Hebei (BTH) region and traffic Beijing year 2016. The resulting WRF-Chem simulated compared against ground-based satellite observations. Our estimated annual BTH emissions (625 Gg) Beijing’s (7.8 within ranges published inventories....
Analyzing graph properties of neural networks has recently gained much attention in attempts to understand how information is processed the brain. Using in-vitro techniques form increased popularity as it allows one develop small, easy record that maintain many larger brain [1]. One widely recognized tool for studying vitro Microelectrode Array (MEAs) on which neurons can be cultured and recorded simultaneously. MEAs used grow from disassociated cells spontaneously connect create these then...
Semantic segmentation in indoor environments is a crucial task for artificial intelligence-driven visual robotics, enabling pixel-level classification results to facilitate robot path planning. Inspired by the success of multimodal models, we propose an end-to-end semantic model image tasks scenes, which call OIPNet. We design OIP module enhance network’s ability extract global information and enable interaction different directions. have validated on NYUv2 Sun RGB-D datasets, experiments...
Computer vision obtains object and environment information by simulating human visual senses borrowing sensory activity. As one of the main tasks computer vision, image classification can be used not only for face recognition, traffic scene retrieval, automatic photo categorization but also as a theoretical basis target detection segmentation. In this paper, we use existing CNN architecture network-ConvNeXt. By adapting modifying residual connectivity convolutional structure network, achieve...