A5087370566- Image Processing Techniques and Applications
- Advanced Vision and Imaging
- Image and Signal Denoising Methods
- Blind Source Separation Techniques
- Face and Expression Recognition
- Neural Networks and Applications
- Advanced Image Processing Techniques
- Video Surveillance and Tracking Methods
- Optical measurement and interference techniques
- Spectroscopy and Chemometric Analyses
- Remote-Sensing Image Classification
- Advanced Image and Video Retrieval Techniques
- Advanced Image Fusion Techniques
- Robotics and Sensor-Based Localization
- Sparse and Compressive Sensing Techniques
- Gene expression and cancer classification
- Machine Learning and ELM
- 3D Shape Modeling and Analysis
- Face recognition and analysis
- Image Enhancement Techniques
- Marine and coastal ecosystems
- Image Retrieval and Classification Techniques
- Marine and fisheries research
- Medical Image Segmentation Techniques
- Speech and Audio Processing
Jiangsu University of Science and Technology
2024-2025
Ocean University of China
2025
Anhui University
2015-2024
Hong Kong Polytechnic University
2007-2020
Leibniz Institute of Agricultural Development in Transition Economies
2020
ORCID
2020
Chinese Academy of Sciences
2004-2016
South China Sea Institute Of Oceanology
2011-2016
Institute of Oceanology
2011-2016
University of Chinese Academy of Sciences
2014
This paper investigates the feasibility of applying a relatively novel neural network technique, i.e., extreme learning machine (ELM), to realize neuro-fuzzy Takagi-Sugeno-Kang (TSK) fuzzy inference system. The proposed method is an improved version regular TSK For method, first, data that are processed grouped by k-means clustering method. membership arbitrary input for each rule then derived through ELM, followed normalization At same time, consequent part rules obtained multiple ELMs....
Differential coexpression analysis usually requires the definition of 'distance' or 'similarity' between measured datasets. Until now, most common choice is Pearson correlation coefficient. However, coefficient sensitive to outliers. Biweight midcorrelation considered be a good alternative since it more robust In this paper, we introduce use Midcorrelation measure gene expression profiles, and provide new approach for differential analysis. Firstly, calculate biweight coefficients all pairs....
3D-printed short-carbon-fiber-reinforced thermoplastic composites have attracted significant attention from both the academic and industrial communities due to their remarkable advantages such as lightweight, high strength, recyclability. However, in most of current 3D-printing-related nylon composites, content short carbon fibers is generally low, influence laws on mechanical properties not been fully explored. This paper focuses (SCF/PA) with short-carbon-fiber contents 15 wt%, 25 35...
In this paper, we propose an efficient algorithm to reconstruct the 3D structure of a human face from one or more its 2D images with different poses. our algorithm, nonlinear least-squares model is first employed estimate depth values facial feature points and pose image concerned by means similarity transform. Furthermore, optimization schemes are presented regard accuracy levels training time required. Our also embeds symmetrical property into procedure, in order alleviate sensitivities...
In this paper, we propose a novel and efficient algorithm to reconstruct the 3-D structure of human face from one or number its 2-D images with different poses. our proposed algorithm, rotation translation process frontal-view image nonfrontal-view is at first formulated as constrained independent component analysis (cICA) model. Then, overcomplete ICA problem converted into normal by incorporating prior CANDIDE Furthermore, model employed construct reference signal that used in both...
How to achieve competitive accuracy and less computation time simultaneously for background estimation is still an intractable task. In this paper, effective subtraction approach video sequences proposed based on a subsuperpixel model. our algorithm, the superpixels of first frame are constructed using simple linear iterative clustering method. After transforming from color format gray level, initial divided into K smaller units, i.e., subsuperpixels, via k-means algorithm. The model then...
The one-sample-per-person problem has become an active research topic for face recognition in recent years because of its challenges and significance real-world applications. However, achieving relatively higher accuracy is still a difficult due to, usually, too few training samples being available variations illumination expression. To alleviate the negative effects caused by these unfavorable factors, this paper we propose more accurate spectral feature image-based 2DLDA (two-dimensional...
Abstract DNA N6-methyladenine (6mA) is produced by the N6 position of adenine being methylated, which occurs at molecular level, and involved in numerous vital biological processes rice genome. Given shortcomings experiments, researchers have developed many computational methods to predict 6mA sites achieved good performance. However, existing do not consider occurrence mechanism extract features from structure. In this paper, a novel deep learning method proposed devising graph feature...
Medium Resolution Imaging Spectrometer (MERIS), Moderate Spectroradiometer (MODIS), and Sea-viewing Wide Field-of-view Sensor (SeaWiFS) particulate organic carbon (POC) concentration products for the South China Sea (SCS) were compared with in situ data collected from October 2007 to December 2013. Spectral remote-sensing reflectance (Rrs,λ) was also measured help understand POC algorithm performance. A strict comparison of satellite-derived measurements showed that MERIS, MODIS, SeaWiFS...
In this letter, a standard postnonlinear blind source separation algorithm is proposed, based on the MISEP method, which widely used in linear and nonlinear independent component analysis. To best suit wide class of mixtures, we adapt method to incorporate priori information mixtures. particular, group three-layered perceptrons network are as unmixing system separate sources another perceptron auxiliary network. The learning for then obtained by maximizing output entropy proposed applied...
Visual pattern of human iris provides rich texture information for personal identification. However, it is challenging to match intra-class images with large variations in applications. This study proposes a perturbation-enhanced feature correlation filter (PFCF) robust matching. PFCF developed based on quad-phase minimum average energy filter, but has two significant improvements. First, performed Gabor filtered encode both local and global features. On the one hand, can enhance details...
The Sea-viewing Wide Field-of-view Sensor (SeaWiFS), Moderate Resolution Imaging Spectroradiometer (MODIS), and Medium Spectrometer (MERIS) remote-sensing radiometric chlorophyll-a (chl-a) concentration products for the South China Sea (SCS) from October 2003 to May 2010 were assessed using in situ data. A strict spatiotemporal match-up method was used minimize temporal variability effects of atmosphere seawater around measurement site. comparison reflectance (Rrs(λ)) three sensors with...
Different substrate types, such as seagrass, sand, stone and silt, can be identified by detecting reflectivity. In this study, high-resolution remote sensing data of Quickbird were used to map seagrass distribution in the northeast coast Xincun Bay. Radiance transfer model optical shallow water was retrieve bottom reflectivity; relationship between hyperspectral band leaf area index (LAI) for processing satellite density submerged seagrass. Hyperspectral results showed that spectral bands at...