Xun Wang’s research while affiliated with Zhejiang Gongshang University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (30)


A Preliminary Geometric Structure Simplification for Principal Component Analysis
  • Article

November 2018

·

10 Reads

·

1 Citation

Neurocomputing

Huamao Gu

·

·

Xun Wang

Real world data are commonly geometrically nonlinear and thus are not easy to be processed by the traditional linear methods. Many existing techniques for nonlinear dimensionality reduction need careful parameter tuning and cannot be applied to real data stably and consistently. In this article we propose an efficient data preprocessing algorithm, called Curve Straightening Transformation (CST), to flatten the nonlinear geometric structure of data. Then Principal Component Analysis (PCA) and other linear projection methods are adequate to perform the dimensionality reduction task in most cases. In this aspect, the proposed CST algorithm can be regarded as a geometric preprocessing step tailored for PCA. The comprehensive experiments on both artificial and real datasets demonstrate that the proposed preprocessing algorithm is able to simplify the nonlinear geometric structures, and the flattened data are suitable for further dimensionality reduction by linear methods such as PCA.


Manifold Learning by Curved Cosine Mapping

July 2017

·

39 Reads

·

5 Citations

IEEE Transactions on Knowledge and Data Engineering

Huamao Gu

·

Xun Wang

·

Xuewen Chen

·

[...]

·

Jinqin Shi

In the field of pattern recognition, data analysis and machine learning, data points are usually modeled as high-dimensional vectors. Due to the curse-of-dimensionality, it is non-trivial to efficiently process the orginal data directly. Given the unique properties of nonlinear dimensionality reduction techniques, nonlinear learning methods are widely adopted to reduce the dimension of data. However, existing nonlinear learning methods fail in many real applications because of the too-strict requirements (for real data) or the difficulty in parameters tuning. Therefore, in this paper, we investigate the manifold learning methods which belong to the family of nonlinear dimensionality reduction methods. Specifically, we proposed a new manifold learning principle for dimensionality reduction named Curved Cosine Mapping (CCM). Based on the law of cosines in Euclidean space, CCM applies a brand new mapping pattern to manifold learning. In CCM, the nonlinear geometric relationships are obtained by utlizing the law of cosines, and then quantified as the dimensionality-reduced features. Compared with the existing approaches, the model has weaker theoretical assumptions over the input data. Moreover, to further reduce the computation cost, an optimized version of CCM is developed. Finally, we conduct extensive experiments over both artificial and real-world datasets to demonstrate the performance of proposed techniques.


DCT-Based Blind Watermarking of 3D Models

December 2013

·

11 Reads

·

1 Citation

Lecture Notes in Electrical Engineering

To propose robust blind watermarking methods of 3D models based on DCT. First a 3D mesh models will be mapped to a 2D parametric mesh with a kind of planar parameterization method, geometric signals are then transformed into 2D signals. Then a DCT-based watermark scheme is proposed to embed the watermark into some DCT coefficients. The watermark can be detected without the original 3D models. Experimental results show that the embedded watermark is robust against various geometry signal processing.


3D Models Simplification Algorithm for Mobile Devices

November 2012

·

17 Reads

·

2 Citations

Journal of Software

As we know, the mobile device screen is small. The lower accuracy of the model is relatively weak and the capacity to handle high detailed model is very limited. What's more, the existing three-dimensional simplification algorithms are for the personal computer and they are not suitable for the mobile terminals. Thus, we propose a novel 3D model simplification algorithm based on feature points. We will find the model surface curvature of each vertex, based on the model feature points. For these points, we present a new method to obtain it. Experimental results show that the algorithm can accurately reflect the model changes of the local surface geometry, and effectively keep the details of the model characteristics.


A Robust Zero-Watermarking Algorithm for Vector Digital Maps Based on Statistical Characteristics

October 2012

·

46 Reads

·

23 Citations

Journal of Software

This paper presented a new zero-watermarking algorithm for vector digital maps based on statistical characteristics. The watermark information is constructed by utilizing the original data's characteristics. We divide the map into rings by using concentric circles and count the number of vertices in each ring, which is the feature information. A zero watermark image is constructed by using feature information and copyright information. Experiments show that the watermarks are resilient to translation, scaling, vertex deletion and growth, rotation, random noise, objects scrambling and cropping, making it a robust algorithm.


Robust 3D Mesh Watermarking Based on Planar Parameterization and DCT

July 2012

·

17 Reads

Journal of Convergence Information Technology

This paper presents a robust 3D mesh watermarking method based on planar parameterization and DCT. Firstly, a planar parameterization algorithm is used to map 3D mesh models to 2D parametric meshes. Geometric signals are then transformed into 2D signals. Then a DCT-based watermark scheme is proposed to embed the watermark into some DCT coefficients. The algorithm adopts a nonblind extraction process to retrieve the watermark. Experiments show that this method is robust against common attacks such as adding noise attacks, uniform affine transformation attacks.


Visual Important Driven Texture Map Selective Compression

April 2012

·

7 Reads

Journal of Computational and Theoretical Nanoscience

In this paper, we present a visual important driven selective compression method for texture map image. It is based on the notion that the visual important area of texture map image is regarded as texel region of interest, which will be compressed less than the other background areas. In order to obtain the visual important areas, we take not only the saliency information of the texture image but also the distortion of texture mapping into accounts that are dependent on the mapped 3D model. With the help of this information, a visual important map is generated. Guided by this map, the texture map image is divided into several discrete texel region of interest and background with different priority. A selective compression method is proposed to compress these texel regions of interest and background with different compression quality by shifting the wavelet coefficients that belong to different subbands. For finding the wavelet coefficients pertaining to these texel regions of interest exactly, a novel mask generation method is presented. Experimental result shows a better performance for the visual important area, if this selective coding method is employed into the whole texture map image.


Deciding the SHOQ(D)-Satisfiability with a Fully Tiered Clause Group

February 2012

·

35 Reads

Journal of Software

SHOQ(D) is one of the fundamental theories In Description Logics due to its support to concrete datatypes and named individuals. At present, deciding the satisfiability of SHOQ(D)-concepts is mainly completed by enhancing Tableau algorithm with blocking. However, there is still much to be desired in performance as there are tremendous description overlaps in completion forest, thus causing great spatial waste as a result. To tackle this problem, this paper presented a new approach to check the satisfiability of acyclic SHOQ(D)-concepts——FTC(Fully Tiered Clauses) algorithm. This calculus can make a direct judgement on the satisfiability of acyclic SHOQ(D)-concept by translating its description into a fully tiered clause group whose satisfiability is directly available, and reusing clauses to block unnecessary extensions. FTC algorithm eliminates description overlaps to the largest extent as it works on concept description directly. Therefore, FTC algorithm has notably better performance than Tableau by saving a lot of spatial costs.


Efficient Panorama Mosaicing Based on Enhanced-FAST and Graph Cuts

January 2012

·

14 Reads

·

3 Citations

Lecture Notes in Electrical Engineering

This paper presents an efficient and accurate method for creating full view panoramas. A new feature points detection algorithm called Enhanced- FAST is proposed to accurately align images and a graph cuts algorithm is used to merge two adjacent images seamlessly. Based on the FAST algorithm, the Enhanced-FAST algorithm smoothes and extends the sampling area making the feature points detection more insensitive to noise. Our graph cuts algorithm uses image Laplacian to compute the edge weights which can find an optimized seam even under different lighting. Experiments and comparisons show that our method is efficient and robust to image noise and lighting changing.


A Mixture of Gaussian-Based Method for Detecting Foreground Object in Video Surveillance

January 2012

·

5 Reads

Lecture Notes in Electrical Engineering

This chapter presents a novel MoG based method for foreground detection and segmentation in video surveillance. Normal MoG is difficult to deal with the foreground objects stay in the scene for a long time and segment different foreground objects from one blob. We improve MoG by adopting posterior feedback information of object tracking based on statistics to robustly modeling the background and to perfect the foreground segmentation result. Experiments and comparisons show that our method is robust and accurate for detecting foreground object in video surveillance.


Citations (11)


... The precision of hyperspectral inversion for the soil Cu content is influenced not only by the spectral preprocessing but also by dimensionality reduction method [56][57][58][59][60][61][62]. A suitable combination of spectral preprocessing and dimensionality reduction can improve the accuracy and performance of the inversion model. ...

Reference:

On Optimizing Hyperspectral Inversion of Soil Copper Content by Kernel Principal Component Analysis
A Preliminary Geometric Structure Simplification for Principal Component Analysis
  • Citing Article
  • November 2018

Neurocomputing

... For data with nonlinear structures, non-linear subspace learning algorithms are required, such as manifold learning methods [7,13] or kernel based subspace learning algorithms [8,14]. Manifold learning methods assume that the highdimensional data lie on a low-dimensional manifold, and explore the local or global data manifold structure information by manifold assumption Kernel based subspace learning methods [5,6] map high-dimensional data to a Reproducing Kernel Hilbert Space (RKHS) through kernel function to reduce the nonlinearity of data, and learning a linear subspace learning in RKHS. ...

Manifold Learning by Curved Cosine Mapping
  • Citing Article
  • July 2017

IEEE Transactions on Knowledge and Data Engineering

... The most classic approach is to detect and extract image point features corresponding to unique landmarks in the scene and then match them across different views. This feature-based mosaicking approach (Milgram, 1975) has been investigated extensively in recent decades, using different wellknown hand-crafted feature approaches such as Harris (Okumura et al., 2013), SIFT (Li et al., 2008), SURF (Rong et al., 2009), ORB (Chaudhari et al., 2017), and FAST (Wang et al., 2012). More recently, datadriven features that are learned by deep neural networks have been utilised for image mosaicking (Bano et al., 2020;Zhang et al., 2019). ...

Efficient Panorama Mosaicing Based on Enhanced-FAST and Graph Cuts
  • Citing Article
  • January 2012

Lecture Notes in Electrical Engineering

... Zhang et al. utilized the coordinate difference between each two points in the model to construct a difference histogram, further improving the embedding capacity of the algorithm [2]. Zhou et al. proposed a 3D model transformation domain information hiding algorithm based on DCT [3]. Ren et al. pro-posed a 3D model information hiding algorithm based on distance features and volume integral invariants, which effectively improved the robustness of the algorithm while ensuring invisibility [4]. ...

DCT-Based Blind Watermarking of 3D Models
  • Citing Article
  • December 2013

Lecture Notes in Electrical Engineering

... The embedding is based on the principle of finding redundant space in the data so that the watermark can be inserted without or with little effect on the data to be protected. The techniques for inserting or hiding this information into other information mainly used today include spectral modulation, least significant bit, or quantization index modulation techniques [27]. Note also that the watermark does not have to be embedded in the carrier data, and techniques of this kind are called zero-watermarking. ...

A Robust Zero-Watermarking Algorithm for Vector Digital Maps Based on Statistical Characteristics
  • Citing Article
  • October 2012

Journal of Software

... Based on the embedding domains as well as the accuracy requirements of the watermarked data, watermarking techniques are divided into several categories such as spatial domain-based watermarking, frequency domainbased watermarking, reversible watermarking, lossless watermarking, and zero-watermarking. Research on watermarking using techniques of embedding information into the spatial domain (coordinates of objects) can be mentioned as [3][4][5][6][7][8][9], techniques of embedding information into the frequency domain such as [10][11][12]. Reversible watermarking techniques for vector geographic data are performed in [13][14][15][16][17][18]. In [19,20], the studies of watermarking without losing information based on the storage order are presented. ...

A DCT-based Blind Watermarking Algorithm for Vector Digital Maps
  • Citing Article
  • January 2011

Advanced Materials Research

... La utilización de formatos de archivo tradicionales orientados a las representaciones de geometría para fines de visualización incluye una pérdida dramática de la semántica, reduciendo toda la información disponible a objetos geométricos sin distinción semántica, como pueden ser los conjuntos de triángulos que representan a un árbol o un edificio. Ling et al. [70] introdujeron una definición conceptual y una implementación preliminar de una capa middleware para hacer frente a este tipo de datos GIS heterogéneos, dados en una serie de formatos de archivo distintos. Sin embargo, la preservación semántica no está realmente garantizada, ya que sólo se considera un método de conversión entre dos formatos de archivo. ...

Integration of Heterogeneous Geospatial Data Based on Middleware Technology
  • Citing Article
  • July 2009

... 38 Wang et al proposed a DCT-based method to put the watermark in the middle frequency coefficients and make the watermark insertion process nearly lossless. 39 Lei et al embedded the watermark in the singular values of low-frequency coefficients based on QWT to improve watermark robustness. 19 As we all know, the high-frequency coefficients correspond to detailed information, while the low-frequency coefficients occupy most of the energy of the image and correspond to large-scale features; therefore, the latter is more stable and is less disturbed by external factors. ...

A remote sensing image self-adaptive blind watermarking algorithm based on wavelet transformation
  • Citing Article
  • September 2007

... In this case, the design as well as the balance of gameplay and education are vital in order to keep the players motivated for playing and learning. In order to make full use of edutainment, learning and gaming should be blended harmoniously [72]. On the one hand, this means to involve teachers in the design of educational games. ...

An Ontology-Based Development Framework for Edutainments
  • Citing Article
  • December 2008