#### DMCA

## Walking (1683)

### Citations

3793 |
Density Estimation for Statistics and Data Analysis, Chapman and
- Silverman
- 1986
(Show Context)
Citation Context ...ints to normalize the density for the purpose of directly comparing densities at location x. A Gaussian kernel is used in our approach with the bandwidth h determined by the Silverman’s rule of thumb =-=[38]-=- . 3.2 Feature-preserving Point Resampling Point sampling is a fundamental tool to reduce the number of data points while best preserving the features for many computational tasks. Due to the uniformi... |

1742 |
The Visual Display of Quantitative Information
- Tufte
- 1983
(Show Context)
Citation Context ... 8 (d)) are put on the left side. 4.2 Visual Exploration of A Single Class To gain an overview of each class, we employ a data class list view (Figure 8 (a)) that follows a small multiple visual form =-=[39]-=-. Specifically, the visual representation of each data class is designed as a rectangular glyph (Figure 7) which contains two statistical histograms. Other information about the data class is also emb... |

985 |
UCI machine learning repository
- Asuncion, Newman
- 2007
(Show Context)
Citation Context ...x-coord of the tag (d) Point counts in thesselected region of (d)s# (s) = 4 # (s) = 27 # (s) = 41 y-co o rdso f th e ta g y-co o rdso f th e ta g Fig. 1. Visualization for the Person Activity dataset =-=[1]-=-. (a) The conventional scatterplot suffers from severe overdraw. (b) The Splatterplots [33] may synthesize new colors. Data points in sparse regions are explicitly highlighted. (c) The Splatterplots w... |

597 | Convergence properties of the Nelder–Mead simplex method in low dimensions
- Lagarias, Reeds, et al.
- 1998
(Show Context)
Citation Context ..., the optimization process minimize Equation (5) by search other two channels (i.e. a* and b*) in an isolightness plane within the gamut of the CIELAB color space. We apply the Nelder and Mead method =-=[29]-=- to find an optimal solution. As the downhill simplex method works iteratively and is prone to converge to a local minima, we run the optimization method several times with random initializations and ... |

158 |
Spot noise - texture synthesis for data visualization
- Wijk
- 1991
(Show Context)
Citation Context ...tanding in meaningful ways. 2.3 Noise for Visualization Noise has been widely used in various visualization techniques. A well-known application is noise-based flow visualization [23, 37]. Spot noise =-=[40]-=- was introduced to vector field visualization by Jark van Wijk. Inspired by this method, a white noise texture was employed by line integral convolution (LIC) based methods [37]. Khlebnikov et al. [28... |

114 | Rolling the dice: Multidimensional visual exploration using scatterplot matrix navigation
- Elmqvist, Dragicevic, et al.
- 2008
(Show Context)
Citation Context ...-semantic zoom that provides a smooth transition between overview and detail, and a fisheye distortion that displays the focus and context regions of the scatterplot in a single view. Elmqvist et al. =-=[17]-=- exploited animated transitions between scatterplots for visual exploration of a multidimensional dataset. Yuan et al. [47] presented a comprehensive tool equipped with selection, zooming, dragging, a... |

86 | Bubble Sets: Revealing set relations with isocontours over existing visualizations.
- Collins, Penn, et al.
- 2009
(Show Context)
Citation Context ...ltidimensional dataset. Yuan et al. [47] presented a comprehensive tool equipped with selection, zooming, dragging, and linking for interactive subspace exploration of 2D scatterplots. Collins et al. =-=[13]-=- enhanced the set relation of points on the plot with clustering and contours. Chan et al. [8, 9] introduced the flow-based scatterplots for sensitivity and local trends analysis where a local regress... |

72 |
Scatterplot Matrix Techniques for Large N”,
- Carr, Littlefield, et al.
- 1987
(Show Context)
Citation Context ...the number of data points falling into them. Alpha blended scatterplots are an example of this technique. This simple discrete form of density is prone to introduce bias in visualization. Carr et al. =-=[7]-=- used hexagonal cells to accumulate densities. Bachthaler et al. [2, 3] proposed a rigorous, accurate, and generic mathematical model to create the continuous scatterplots. Zinsmaier et al. [48] emplo... |

57 | Uflic—a Line Integral Convolution Algorithm for Visualizing Unsteady Flows
- Shen, Kao
- 1997
(Show Context)
Citation Context ...enhance users’ understanding in meaningful ways. 2.3 Noise for Visualization Noise has been widely used in various visualization techniques. A well-known application is noise-based flow visualization =-=[23, 37]-=-. Spot noise [40] was introduced to vector field visualization by Jark van Wijk. Inspired by this method, a white noise texture was employed by line integral convolution (LIC) based methods [37]. Khle... |

33 | Continuous scatterplots
- Bachthaler, Weiskopf
- 2008
(Show Context)
Citation Context ...fects. In general, one can modulate the limited visual channels of points, e.g., the point size [30, 31] or the opacity [18] to decrease the overplotting degree. Alternatively, the density estimation =-=[2, 3, 18]-=- can be employed to reformulate the point distribution in a simple form. This is effective for clarifying dense regions, but can hardly show outliers which typically lie in low density regions. Spatia... |

33 | High-dimensional visual analytics: Interactive exploration guided by pairwise views of point distributions.
- Wilkinson, Anand, et al.
- 2006
(Show Context)
Citation Context ...he plot with clustering and contours. Chan et al. [8, 9] introduced the flow-based scatterplots for sensitivity and local trends analysis where a local regression analysis is leveraged. Leland et al. =-=[15, 42]-=- defined a set of measures to characterize the scatterplot point distributions and used these measures to organize scatterplots for high-dimensional data analysis. Radloff et al. [34] used additional ... |

30 |
Visualization of vessel movements
- Willems, Wetering, et al.
- 2009
(Show Context)
Citation Context ...maier et al. [48] employed the kernel density estimation (KDE) to visualize overlapped nodes and edges in large graphs. The density estimation can also be adopted to visualize overlapped trajectories =-=[36, 43]-=-. Color mapping and contouring are two common methods to visualize and highlight dense regions in a reconstructed continuous density field. Unfortunately, these methods neglect the outliers in the res... |

29 |
The Gridfit Algorithm: An Efficient and Effective Approach to Visualizing Large Amounts of Spatial Data.
- Keim, Herrmann
- 1998
(Show Context)
Citation Context ...in a simple form. This is effective for clarifying dense regions, but can hardly show outliers which typically lie in low density regions. Spatially moving the overplotted points to unoccupied pixels =-=[25, 26]-=- is another way to address the overdraw problem. However, it is not suitManuscript received 31 Mar. 2014; accepted 1 Aug. 2014 ate of publication 2014; date of current version 2014.11sAug.s9sNov.sD.sD... |

28 | Constant density visualizations of non-uniform distributions of data
- Woodruff, Landay, et al.
(Show Context)
Citation Context ...he overdraw problem. Conventional solutions can be roughly categorized into three classes. Changing the visual channels are a simple and intuitive way to deal with the overdraw issue. Woodruff et al. =-=[44]-=- chose to visualize scatter dataset with small-size dots in highly dense regions and large-size icons in sparse regions for outliers, as small dots occupy relatively less screen space. A similar schem... |

27 | User interaction with scatterplots on small screens—a comparative evaluation of geometric-semantic zoom and fisheye distortion
- Bu¨ring, Gerken, et al.
- 2006
(Show Context)
Citation Context ...tive Exploration and Analysis for Scatterplots A bunch of navigation tools and scatterplot variants were developed to assist users discovering and exploring the insights from datasets. Buering et al. =-=[5]-=- proposed two interaction techniques to explore a large dataset in a small screen: a geometric-semantic zoom that provides a smooth transition between overview and detail, and a fisheye distortion tha... |

26 | Anisotropic blue noise sampling
- LI, WEI, et al.
- 2010
(Show Context)
Citation Context ...a reconstruction-and-resampling process that generates a visual abstraction of the input point distribution. To make it amenable for multi-class points, a multi-class blue noise based sampling scheme =-=[41]-=- is employed. We also propose a color optimization technique to enhance the perceptual contrast in multi-class regions. The integrated visual exploration system consists of a suite of distribution stu... |

23 |
Weaving versus blending: a quantitative assessment of the information carrying capacities of two alternative methods for conveying multivariate data with color.
- Hagh-Shenas, Interrante, et al.
- 2006
(Show Context)
Citation Context ...l scatterplot suffers from severe overdraw. (b) The Splatterplots [33] may synthesize new colors. Data points in sparse regions are explicitly highlighted. (c) The Splatterplots with additional noise =-=[22]-=- can enhance the distinguishability in dense regions. (d) Our result: please notice the differences in overlapped regions indicated by the circles. The number of data points before and after abstracti... |

20 | A novel interface for interactive exploration of DTI fibers
- Chen, Ding, et al.
- 2009
(Show Context)
Citation Context ...tations, scatterplots use 2D Cartesian coordinates to depict a set of bivariate points. The point collection in a 2D plane can be explored and analyzed to study point distributions [15, 42], clusters =-=[11, 47]-=-, outliers [33], local trends [8, 9], and axial correlations [25]. One challenging problem for visualizing scatterplots is the overdraw (see Figure 1 (a)) caused by a dense point distribution. In the ... |

19 |
Wysiwyg (what you see is what you get) volume visualization
- Guo, Mao, et al.
- 2011
(Show Context)
Citation Context ...f the selected data classes. It is of equal, if not more, importance to interactively explore the data from different perspectives. Our system provides a set of tools inspired by many editing systems =-=[6, 21]-=-. 4.3.1 Data Inspection Users are allowed to use the following tools to study multi-class scatter dataset in a single view: CHEN ET AL.: VISUAL ABSTRACTION AND EXPLORATION OF MULTI-CLASS SCATTERPLOTS ... |

16 |
Matching visual saliency to confidence in plots of uncertain data
- Feng, Kwock, et al.
(Show Context)
Citation Context ...ade, many research efforts have concentrated on reduction of overdraw and its effects. In general, one can modulate the limited visual channels of points, e.g., the point size [30, 31] or the opacity =-=[18]-=- to decrease the overplotting degree. Alternatively, the density estimation [2, 3, 18] can be employed to reformulate the point distribution in a simple form. This is effective for clarifying dense re... |

15 | Interactive level-of-detail rendering of large graphs.
- Zinsmaier, Brandes, et al.
- 2012
(Show Context)
Citation Context ...et al. [7] used hexagonal cells to accumulate densities. Bachthaler et al. [2, 3] proposed a rigorous, accurate, and generic mathematical model to create the continuous scatterplots. Zinsmaier et al. =-=[48]-=- employed the kernel density estimation (KDE) to visualize overlapped nodes and edges in large graphs. The density estimation can also be adopted to visualize overlapped trajectories [36, 43]. Color m... |

14 |
Composite density maps for multivariate trajectories
- Scheepens, Willems, et al.
(Show Context)
Citation Context ...maier et al. [48] employed the kernel density estimation (KDE) to visualize overlapped nodes and edges in large graphs. The density estimation can also be adopted to visualize overlapped trajectories =-=[36, 43]-=-. Color mapping and contouring are two common methods to visualize and highlight dense regions in a reconstructed continuous density field. Unfortunately, these methods neglect the outliers in the res... |

13 |
Give chance a chance: modeling density to enhance scatter plot quality through random data sampling
- Bertini, Santucci
(Show Context)
Citation Context ...y extended this method to multivariate volume data visualization [27]. A perceptually adapted Perlin Noise was employed by Coninx et al. [14] to visualize uncertain scalar field. Bertini and Santucci =-=[4]-=- presented a random data sampling method to model the underlying data density for scatterplot quality enhancement. Unfortunately it is intractable for multi-class scatter datasets. Our work leverages ... |

12 | Wang W.: Variational blue noise sampling
- Chen, Yuan, et al.
- 2012
(Show Context)
Citation Context ...Resampling Point sampling is a fundamental tool to reduce the number of data points while best preserving the features for many computational tasks. Due to the uniformity and absence of spectral bias =-=[10]-=-, the blue noise sampling technique has been favored in many applications such as image stippling [41, 46] and point cloud resampling [10]. Individually sampling each data class cannot guarantee that ... |

12 |
Generalized scatter plots.
- Keim, Hao, et al.
- 2010
(Show Context)
Citation Context ... bivariate points. The point collection in a 2D plane can be explored and analyzed to study point distributions [15, 42], clusters [11, 47], outliers [33], local trends [8, 9], and axial correlations =-=[25]-=-. One challenging problem for visualizing scatterplots is the overdraw (see Figure 1 (a)) caused by a dense point distribution. In the past decade, many research efforts have concentrated on reduction... |

11 | Efficient and adaptive rendering of 2D continuous scatterplots
- Bachthaler, Weiskopf
(Show Context)
Citation Context ...fects. In general, one can modulate the limited visual channels of points, e.g., the point size [30, 31] or the opacity [18] to decrease the overplotting degree. Alternatively, the density estimation =-=[2, 3, 18]-=- can be employed to reformulate the point distribution in a simple form. This is effective for clarifying dense regions, but can hardly show outliers which typically lie in low density regions. Spatia... |

11 | Direct volume editing
- Burger, Kruger, et al.
- 2008
(Show Context)
Citation Context ...f the selected data classes. It is of equal, if not more, importance to interactively explore the data from different perspectives. Our system provides a set of tools inspired by many editing systems =-=[6, 21]-=-. 4.3.1 Data Inspection Users are allowed to use the following tools to study multi-class scatter dataset in a single view: CHEN ET AL.: VISUAL ABSTRACTION AND EXPLORATION OF MULTI-CLASS SCATTERPLOTS ... |

11 |
A.: Stacking graphic elements to avoid overplotting
- Dang, Wilkinson, et al.
- 2010
(Show Context)
Citation Context ...scheme was exploited by Chen et al. [11] to explore the resulting projection for DTI fibers. The Utopian [12] system generated an alpha blended scatterplot for interactive topic modeling. Dang et al. =-=[16]-=- proposed a method to handle overplotting issue by stacking points in the third dimension. Luboschik et al. [32] introduced the aligned weaving technique that can improve the representation of overlap... |

10 | Flow-based scatterplots for sensitivity analysis
- Chan, Correa, et al.
- 2010
(Show Context)
Citation Context ... coordinates to depict a set of bivariate points. The point collection in a 2D plane can be explored and analyzed to study point distributions [15, 42], clusters [11, 47], outliers [33], local trends =-=[8, 9]-=-, and axial correlations [25]. One challenging problem for visualizing scatterplots is the overdraw (see Figure 1 (a)) caused by a dense point distribution. In the past decade, many research efforts h... |

10 | H.: Utopian: User-driven topic modeling based on interactive nonnegative matrix factorization
- Choo, Lee, et al.
- 1992
(Show Context)
Citation Context ...s in sparse regions for outliers, as small dots occupy relatively less screen space. A similar scheme was exploited by Chen et al. [11] to explore the resulting projection for DTI fibers. The Utopian =-=[12]-=- system generated an alpha blended scatterplot for interactive topic modeling. Dang et al. [16] proposed a method to handle overplotting issue by stacking points in the third dimension. Luboschik et a... |

9 | Dimension projection matrix/tree: Interactive subspace visual exploration and analysis of high dimensional data
- Yuan, Ren, et al.
(Show Context)
Citation Context ...tations, scatterplots use 2D Cartesian coordinates to depict a set of bivariate points. The point collection in a 2D plane can be explored and analyzed to study point distributions [15, 42], clusters =-=[11, 47]-=-, outliers [33], local trends [8, 9], and axial correlations [25]. One challenging problem for visualizing scatterplots is the overdraw (see Figure 1 (a)) caused by a dense point distribution. In the ... |

8 |
Visualization of uncertain scalar data fields using color scales and perceptually adapted noise
- Coninx, Bonneau, et al.
- 2011
(Show Context)
Citation Context ...ing Gabor noise to visualize 2D multivariate dataset. Later they extended this method to multivariate volume data visualization [27]. A perceptually adapted Perlin Noise was employed by Coninx et al. =-=[14]-=- to visualize uncertain scalar field. Bertini and Santucci [4] presented a random data sampling method to model the underlying data density for scatterplot quality enhancement. Unfortunately it is int... |

8 | A new weaving technique for handling overlapping regions
- Luboschik, Radloff, et al.
(Show Context)
Citation Context ...stem generated an alpha blended scatterplot for interactive topic modeling. Dang et al. [16] proposed a method to handle overplotting issue by stacking points in the third dimension. Luboschik et al. =-=[32]-=- introduced the aligned weaving technique that can improve the representation of overlapping clusters in scatterplots. Generally these methods are effective, but the number of usable visual channels a... |

7 |
A model of symbol size discrimination in scatterplots
- Li, Martens, et al.
- 2010
(Show Context)
Citation Context ...ibution. In the past decade, many research efforts have concentrated on reduction of overdraw and its effects. In general, one can modulate the limited visual channels of points, e.g., the point size =-=[30, 31]-=- or the opacity [18] to decrease the overplotting degree. Alternatively, the density estimation [2, 3, 18] can be employed to reformulate the point distribution in a simple form. This is effective for... |

6 |
Timeseer: Scagnostics for high-dimensional time series
- Dang, Anand, et al.
(Show Context)
Citation Context ...tal visual representations, scatterplots use 2D Cartesian coordinates to depict a set of bivariate points. The point collection in a 2D plane can be explored and analyzed to study point distributions =-=[15, 42]-=-, clusters [11, 47], outliers [33], local trends [8, 9], and axial correlations [25]. One challenging problem for visualizing scatterplots is the overdraw (see Figure 1 (a)) caused by a dense point di... |

6 | Perception of average value in multiclass scatterplots.
- Gleicher, Correll, et al.
- 2013
(Show Context)
Citation Context ...hierarchical sampling scheme, some points maybe discarded when the view is zoomed in; Bottom: our scheme preserves the points of low levels when the view is zoomed in. 3.4.1 Point Color A recent work =-=[19]-=- suggests that color outperforms shape for multiclass scatterplots in many comparative tasks. Even when the number of points increases, the performance remains good. However, using color to distinguis... |

6 | Paint inspired color mixing and compositing for visualization
- Gossett, Chen
- 2004
(Show Context)
Citation Context ...us density fields with respect to each class of points can be reconstructed. Then the commonly used multi-variate data visualization methods such as contouring, color blending [33], and color weaving =-=[20, 22]-=- are utilized. Nevertheless, these methods may yield misleading representations, especially in regions where multiple classes exist. In addition, the relative density orders among classes cannot be pr... |

5 | The generalized sensitivity scatterplot - Chan, Correa, et al. |

5 | Splatterplots: Overcoming overdraw in scatter plots
- Mayorga, Gleicher
(Show Context)
Citation Context ...s) = 41 y-co o rdso f th e ta g y-co o rdso f th e ta g Fig. 1. Visualization for the Person Activity dataset [1]. (a) The conventional scatterplot suffers from severe overdraw. (b) The Splatterplots =-=[33]-=- may synthesize new colors. Data points in sparse regions are explicitly highlighted. (c) The Splatterplots with additional noise [22] can enhance the distinguishability in dense regions. (d) Our resu... |

4 | Image-space texture-based output-coherent surface flow visualization,” Visualization and Computer Graphics - Huang, Pan, et al. - 2013 |

4 | Visizer: a visualization resizing framework
- Wu, Liu, et al.
(Show Context)
Citation Context ...t, which allows users to strike a balance between overplotting and distortion. Janetzko et al. [24] proposed to enhance the scatterplots using the ellipsoid pixel placement scheme. Further, Wu et al. =-=[45]-=- presented a warping method to avoid important features overplotted during resizing. In essence, distortion based methods change the underlying topology and distribution of the dataset, which may caus... |

3 |
Cartography and the reality of boundaries
- Rankin
(Show Context)
Citation Context ...xample (see the bottom left in Figure 2), the relative population density and transitions among races are missing in each region (e.g., the one indicated by the black arrow). In contrast, Bill Rankin =-=[35]-=- proposed a dot map technique, with which the racial distributions in a region are directly represented by the density of multiple sets of colored dots. The dot color represents the race and the densi... |

2 | Enhancing scatter plots using ellipsoid pixel placement and shading
- Janetzko, Hao, et al.
- 2013
(Show Context)
Citation Context ...ling curve. Later, Keim et al. [25] developed a more flexible technique called the generalized scatterplot, which allows users to strike a balance between overplotting and distortion. Janetzko et al. =-=[24]-=- proposed to enhance the scatterplots using the ellipsoid pixel placement scheme. Further, Wu et al. [45] presented a warping method to avoid important features overplotted during resizing. In essence... |

2 |
Noisebased volume rendering for the visualization of multivariate volumetric data
- Khlebnikov, Kainz, et al.
(Show Context)
Citation Context ... methods [37]. Khlebnikov et al. [28] investigated the possibility of using Gabor noise to visualize 2D multivariate dataset. Later they extended this method to multivariate volume data visualization =-=[27]-=-. A perceptually adapted Perlin Noise was employed by Coninx et al. [14] to visualize uncertain scalar field. Bertini and Santucci [4] presented a random data sampling method to model the underlying d... |

2 |
Procedural texture synthesis for zoom-independent visualization of multivariate data
- Khlebnikov, Kainz, et al.
- 2012
(Show Context)
Citation Context ...40] was introduced to vector field visualization by Jark van Wijk. Inspired by this method, a white noise texture was employed by line integral convolution (LIC) based methods [37]. Khlebnikov et al. =-=[28]-=- investigated the possibility of using Gabor noise to visualize 2D multivariate dataset. Later they extended this method to multivariate volume data visualization [27]. A perceptually adapted Perlin N... |

2 |
Evaluation of symbol contrast in scatterplots
- Li, Wijk, et al.
- 2009
(Show Context)
Citation Context ...ibution. In the past decade, many research efforts have concentrated on reduction of overdraw and its effects. In general, one can modulate the limited visual channels of points, e.g., the point size =-=[30, 31]-=- or the opacity [18] to decrease the overplotting degree. Alternatively, the density estimation [2, 3, 18] can be employed to reformulate the point distribution in a simple form. This is effective for... |

2 | Supporting display scalability by redundant mapping
- Radloff, Luboschik, et al.
- 2011
(Show Context)
Citation Context ...eland et al. [15, 42] defined a set of measures to characterize the scatterplot point distributions and used these measures to organize scatterplots for high-dimensional data analysis. Radloff et al. =-=[34]-=- used additional visual encodings to represent dots in scatterplots on heterogenous displays. In this work, we provide a complementary means to improve the readability of multi-class scatterplots. We ... |

2 | Generating pointillism paintings based on seurat’s color composition
- Wu, Tsai, et al.
- 2013
(Show Context)
Citation Context ...ing the features for many computational tasks. Due to the uniformity and absence of spectral bias [10], the blue noise sampling technique has been favored in many applications such as image stippling =-=[41, 46]-=- and point cloud resampling [10]. Individually sampling each data class cannot guarantee that the blue noise features are preserved for the union of all data classes. Therefore, we employ the adaptive... |