A5084505514- Rough Sets and Fuzzy Logic
- Data Mining Algorithms and Applications
- High-Energy Particle Collisions Research
- Particle physics theoretical and experimental studies
- Text and Document Classification Technologies
- Quantum Chromodynamics and Particle Interactions
- Wireless Signal Modulation Classification
- Image Processing and 3D Reconstruction
- Image and Signal Denoising Methods
- Indoor and Outdoor Localization Technologies
- Imbalanced Data Classification Techniques
- Particle Detector Development and Performance
- Advanced Vision and Imaging
- Underwater Vehicles and Communication Systems
- Robotics and Sensor-Based Localization
- Image Processing Techniques and Applications
- Advanced Steganography and Watermarking Techniques
- Chaos-based Image/Signal Encryption
- Advanced Image Processing Techniques
- Genetics and Neurodevelopmental Disorders
- Blind Source Separation Techniques
- Retinal Imaging and Analysis
- Video Surveillance and Tracking Methods
- Error Correcting Code Techniques
- Advanced Text Analysis Techniques
Southwest Jiaotong University
2014-2024
Zhuhai Institute of Advanced Technology
2021-2024
Jiangxi Agricultural University
2019-2024
Ningxia Medical University
2024
Fujian Polytechnic of Information Technology
2024
Xi'an University of Science and Technology
2021-2023
Peking University
2007-2023
State Key Laboratory of Nuclear Physics and Technology
2023
Chengdu Normal University
2023
University of Science and Technology of China
2023
Feature selection aims to remove irrelevant or redundant features and thereby remain relevant informative so that it is often preferred for alleviating the dimensionality curse, enhancing learning performance, providing better readability interpretability, on. Data contain numerical categorical representations are called heterogeneous data, they exist widely in many real-world applications. Neighborhood rough set (NRS) can effectively deal with data by using neighborhood binary relation,...
Age-related macular degeneration is one of the main causes blindness. However, internal structures retinas are complex and difficult to be recognized due occurrence neovascularization. Traditional surface detection methods may fail in layer segmentation. In this paper, a supervised method reported for simultaneously segmenting layers Three spatial features, seven gray-level-based 14 layer-like features extracted neural network classifier. The coarse surfaces different optical coherence...
Abstract Human ENGRAILED 2 ( EN2 ) gene is localized to 7q36, an autism susceptibility locus. En2 knockout mice display hypoplasia of cerebellum and a decrease in the number Purkinje cell, which are similar those reported for individuals with autism. Furthermore, deficits social behavior were detected −/− mice. Two recent studies have demonstrated that two intronic SNPs (rs1861972, rs1861973) significantly associated To investigate whether this finding could be replicated Chinese Han...
Deep-learning-based modulation recognition methods can extract the features of signals automatically with usage deep neural network (DNN). However, background noises might lower accuracy and induce longer convergence time. In order to improve reduce computational complexity, in this article, we propose construct dataset based on spectral correlation function (SCF), which has special property relatively being insensitive noises. Moreover, add a convolutional denoising module combat for better...
Fuzzy rough set theory can model uncertainty in data and has been applied to feature selection for machine learning tasks. The existence of noise is one the reasons uncertainty. However, most classical fuzzy models are often sensitive data, which somewhat degrades their applicability process data. Furthermore, a robust evaluation function nontrivial as nonoptimal subsets may be selected due perturbations from redundant features. In this article, we delve into local density indispensable...
The noises embedded in signals will degrade the signal processing quality. Traditional denoising algorithms might not work practical systems since statistical characteristics of be learned. To address this issue, we propose an efficient residual shrinkage convolutional neural network (RSCNN) aided denoiser based on principle domain transformation, shrinking and inverse transforming operations conducted by traditional denoiser. proposed RSCNN is composed batch normalization layer,...
Cystoid macular edema (CME) and hole (MH) are the leading causes for visual loss in retinal diseases. The volume of CMEs can be an accurate predictor prognosis. This paper presents automatic method to segment from abnormal retina with coexistence MH three-dimensional-optical coherence tomography images. proposed framework consists preprocessing segmentation. part includes denoising, intraretinal layers segmentation flattening, vessel silhouettes exclusion. In segmentation, a three-step...
Autism spectrum disorder (ASD) is a neurodevelopmental and network-level mainly diagnosed in children. The aim of the current study was to develop computer-aided diagnosis method with high accuracy distinguish school-aged children (5-12 years) ASD from those typically developing (TD). used multi-institutional functional magnetic resonance imaging (fMRI) datasets 198 participants Brain Imaging Data Exchange II database employed enhanced stacked auto-encoders between TD. In study, average...
In coexistent heterogeneous wireless networks, receivers have to process the signaling and data following different specifications. With aim automatically intelligently identify received signal type then recover data, in this paper, we propose a unified intelligent channel decoder serially concatenated by convolutional-neural-network-based classifier deep learning (DL)-aided decoder. The mainly consists of convolutional layer, batch normalization max-pooling while DL is constituted four full...
The digital economy has become an important force driving the transformation of old and new forces in China's economy, also provides opportunity for retail enterprises to "overtake" by changing lanes. evaluation maturity plays role their process. Although more are realizing own development, is a complex issue that involves all aspects enterprise management. There still many lack clear strategic goals practical paths, as well effective supporting assessments institutional incentives process...
Feature extraction is a fundamental and challenging task in machine learning, which aims at extracting subset of significant discriminant features from raw data for various downstream tasks. The process involves mapping the original into space with lower dimensions while preserving desirable information. However, often has hidden manifold structures, contain neighbour sample information within same class. Most existing methods that extract without considering potentially structures would...
The noise existent in practical systems degrades reliability performances, especially when delivering data over fast fading channels with the correlated noise. To address this issue, we propose a residual convolutional neural network aided denoiser (ResCNND) to suppress for more reliable information recovery. intelligent ResCNND is composed by layers, batch normalization layers and nonlinear activation functions. then integrated different types of channel decoders, preprocessing module embed...