A5101944125- Image and Signal Denoising Methods
- Satellite Communication Systems
- Mobile Agent-Based Network Management
- Advanced Image Fusion Techniques
- Wireless Communication Networks Research
- Network Traffic and Congestion Control
- Radiomics and Machine Learning in Medical Imaging
- Distributed and Parallel Computing Systems
- Neural Networks and Applications
- Remote Sensing in Agriculture
- Remote-Sensing Image Classification
- Wireless Signal Modulation Classification
- Spectroscopy and Chemometric Analyses
- Video Analysis and Summarization
- Hydrological Forecasting Using AI
- Advanced Image Processing Techniques
- Sparse and Compressive Sensing Techniques
- COVID-19 diagnosis using AI
- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Internet Traffic Analysis and Secure E-voting
- Visual Attention and Saliency Detection
- Traffic Prediction and Management Techniques
- Time Series Analysis and Forecasting
- Anomaly Detection Techniques and Applications
China Academy of Art
2011-2024
Lanzhou Jiaotong University
2022-2024
Xidian University
2003-2024
Wuhan Institute of Technology
2021-2024
Nanchang Institute of Science & Technology
2022
Naval University of Engineering
2019-2022
The First People's Hospital of Guiyang
2021
Xian Yang Central Hospital
2021
Yunnan University
2020
Chinese Academy of Sciences
2019
The problem of cloud data classification from satellite imagery using neural networks is considered in this paper. Several image transformations such as singular value decomposition (SVD) and wavelet packet (WP) were used to extract the salient spectral textural features attributed both visible infrared (IR) channels. In addition, well-known gray-level cooccurrence matrix (GLCM) method examined for sake comparison. Two different neural-network paradigms namely probability network (PNN)...
Abstract Background This study aimed to examine multi‐dimensional MRI features’ predictability on survival outcome and associations with differentially expressed Genes ( RNA Sequencing) in groups of glioblastoma multiforme GBM ) patients. Methods Radiomics features were extracted from segmented lesions T2‐ FLAIR data 137 include intensity, shape textural seven classes included the analysis. Patients divided into two depending their time (shorter or longer than 1‐year survival). Four...
Edge detection is a boundary-based segmentation method to extract important information from an image, and it research hotspot in the fields of computer vision image analysis. Especially feature extraction also basis segmentation, target detection, recognition. In recent years, order solve problems edge refinement low accuracy, industry has proposed multiscale fusion wavelet edge, spectral clustering, network reconstruction, other algorithms based on deep learning. enable researchers...
Abstract Multi-focus image fusion is a process of fusing multiple images different focus areas into total image, which has important application value. In view the defects current method in detail information retention effect original architecture based on two stages designed. training phase, combined with polarized self-attention module and DenseNet network structure, an encoder-decoder structure designed for reconstruction tasks to enhance ability model. stage, encoded feature map,...
This paper presents a study of applying an image superresolution method to enhance spatial resolution MRI heart images from temporal sequence. The particular is wavelet-based projection-onto-convex-set reconstruction. It makes use the non-stationary effect in successive sequence extract some detail information for reconstruction at higher resolution. Our experimental result shows better visual quality than that obtained by using cubic-spline interpolation.
Radio spectrum resource is a scarce resource. How to use the limited provide users with maximum communication opportunities worth further study. In fact, traditional "command and control" mode of management very inefficient in utilization spectrum, resulting great waste frequency resources. At same time, radio management, it necessary set up databases different services store monitoring information. Although system plays certain control ability, records form an island information,...
Presents a training algorithm for probabilistic neural networks (PNN) using the minimum classification error (MCE) criterion. A comparison is made between MCE scheme and widely used maximum likelihood (ML) learning on cloud problem satellite imagery data.
Personalized adaptive learning provides smart education with technologies and methodological techniques that enable the personalized aspirations of learners. This paper analyses relevant literature elaborates on concept impact learning. The research systems is systematically reviewed interpreted from three aspects: learner models, resource models algorithms for recommendations. general framework first summarized, followed by ideas methods constructing then implementation mechanism explored,...
The end-to-end model is an important research direction in the field of automatic driving. Reinforcement learning a common policy for training models. Because Deep Deterministic Policy Gradient algorithm can solve problem continuous motion space, it good method to deal with However, uses random sampling method, which makes neural network slow. We propose reinforcement combining Prioritized Experience Replay and Gradient, making convergence faster.
Abstract Stitched images can offer a broader field of view, but their boundaries be irregular and unpleasant. To address this issue, current methods for rectangling start by distorting local grids multiple times to obtain rectangular with regular boundaries. However, these result in content distortion missing boundary information. We have developed an image solution using the reparameterized transformer structure, focusing on single distortion. Additionally, we designed assisted learning...
A new subband-based classification scheme is developed for classifying underwater mines and mine-like targets from the acoustic backscattered signals. The system consists of a feature extractor using wavelet packets, selection scheme, backpropagation neural network classifier. data set used signals seven frequency bands six different objects: two four non-targets. are insonified at 72 aspect angles 0 to 355 degrees with 5 degree increment. Simulation results on ten realizations this...
The detection and prediction of sea clutter power is the basis inversing atmospheric duct. At present, technology duct within radar range relatively perfect, but long-distance inversion limited by range, echo measured Based on theory weighted Markov model grey model, a constructed, sliding method introduced to establish model. relative error between predicted values above four models calculated analyzed using experimental data collected. results show that has better accuracy not only in...