Visual Abstraction and Exploration of Multi-class Scatterplots Haidong Chen, Wei Chen, Honghui Mei, Zhiqi Liu, Kun Zhou, Weifeng Chen, Wentao Gu, Kwan-Liu.

Slides:



Advertisements
Similar presentations
Shape Context and Chamfer Matching in Cluttered Scenes
Advertisements

Complexity Metrics for Design & Manufacturability Analysis
ENV 2006 CS2.1 Envisioning Information: Case Study 2 Scatter Plots on Handheld Devices.
An Interactive-Voting Based Map Matching Algorithm
Accelerating Spatially Varying Gaussian Filters Jongmin Baek and David E. Jacobs Stanford University.
Color Compatibility From Large Datasets Peter O’Donovan University of Toronto Aseem Agarwala Adobe Systems, Inc. Aaron Hertzmann University of Toronto.
Rolling the Dice: Multidimensional Visual Exploration using Scatterplot Matrix Navigation 1 Niklas Elmqvist | Purdue University Pierre Dragicevic | INRIA.
Constructing Popular Routes from Uncertain Trajectories Ling-Yin Wei 1, Yu Zheng 2, Wen-Chih Peng 1 1 National Chiao Tung University, Taiwan 2 Microsoft.
Relevance Feedback Content-Based Image Retrieval Using Query Distribution Estimation Based on Maximum Entropy Principle Irwin King and Zhong Jin Nov
A New Biclustering Algorithm for Analyzing Biological Data Prashant Paymal Advisor: Dr. Hesham Ali.
1 Presented by Jean-Daniel Fekete. 2  Motivation  Mélange [Elmqvist 2008] Multiple Focus Regions.
Modeling Pixel Process with Scale Invariant Local Patterns for Background Subtraction in Complex Scenes (CVPR’10) Shengcai Liao, Guoying Zhao, Vili Kellokumpu,
HDDVis: An Interactive Tool for High Dimensional Data Visualization by Mingyue Tan April 21st, 2004.
Overview Of Clustering Techniques D. Gunopulos, UCR.
Face Poser: Interactive Modeling of 3D Facial Expressions Using Model Priors Manfred Lau 1,3 Jinxiang Chai 2 Ying-Qing Xu 3 Heung-Yeung Shum 3 1 Carnegie.
The Table Lens: Merging Graphical and Symbolic Representations in an Interactive Focus + Context Visualization for Tabular Information R. Rao and S. K.
Table Lens From papers 1 and 2 By Tichomir Tenev, Ramana Rao, and Stuart K. Card.
Bootstrapping a Heteroscedastic Regression Model with Application to 3D Rigid Motion Evaluation Bogdan Matei Peter Meer Electrical and Computer Engineering.
1 Motivation Video Communication over Heterogeneous Networks –Diverse client devices –Various network connection bandwidths Limitations of Scalable Video.
CAD/Graphics 2013, Hong Kong An Image-space Energy-saving Visualization Scheme for OLED Displays Haidong Chen 1, Ji Wang 2, Weifeng Chen 3, Huamin Qu 4,
Input: Original intensity image. Target intensity image (i.e. a value sketch). Using Value Images to Adjust Intensity in 3D Renderings and Photographs.
Jinhui Tang †, Shuicheng Yan †, Richang Hong †, Guo-Jun Qi ‡, Tat-Seng Chua † † National University of Singapore ‡ University of Illinois at Urbana-Champaign.
By LaBRI – INRIA Information Visualization Team. Tulip 2010 – version Tulip is an information visualization framework dedicated to the analysis.
Machine Learning CUNY Graduate Center Lecture 3: Linear Regression.
Gwangju Institute of Science and Technology Intelligent Design and Graphics Laboratory Multi-scale tensor voting for feature extraction from unstructured.
CS654: Digital Image Analysis Lecture 3: Data Structure for Image Analysis.
Graphite 2004 Statistical Synthesis of Facial Expressions for the Portrayal of Emotion Lisa Gralewski Bristol University United Kingdom
Light Using Texture Synthesis for Non-Photorealistic Shading from Paint Samples. Christopher D. Kulla, James D. Tucek, Reynold J. Bailey, Cindy M. Grimm.
International Conference on Intelligent and Advanced Systems 2007 Chee-Ming Ting Sh-Hussain Salleh Tian-Swee Tan A. K. Ariff. Jain-De,Lee.
Visualization Blaz Zupan Faculty of Computer & Info Science University of Ljubljana, Slovenia.
Person detection, tracking and human body analysis in multi-camera scenarios Montse Pardàs (UPC) ACV, Bilkent University, MTA-SZTAKI, Technion-ML, University.
Particle Filters for Shape Correspondence Presenter: Jingting Zeng.
Towards Robust Indexing for Ranked Queries Dong Xin, Chen Chen, Jiawei Han Department of Computer Science University of Illinois at Urbana-Champaign VLDB.
Metro Transit-Centric Visualization for City Tour Planning Pio Claudio and Sung-Eui Yoon.
Enhancing Interactive Visual Data Analysis by Statistical Functionality Jürgen Platzer VRVis Research Center Vienna, Austria.
5. SUMMARY & CONCLUSIONS We have presented a coarse to fine minimization framework using a coupled dual ellipse model to form a subspace constraint that.
80 million tiny images: a large dataset for non-parametric object and scene recognition CS 4763 Multimedia Systems Spring 2008.
Exploiting Context Analysis for Combining Multiple Entity Resolution Systems -Ramu Bandaru Zhaoqi Chen Dmitri V.kalashnikov Sharad Mehrotra.
Data Mining – Intro. Course Overview Spatial Databases Temporal and Spatio-Temporal Databases Multimedia Databases Data Mining.
A Novel Local Patch Framework for Fixing Supervised Learning Models Yilei Wang 1, Bingzheng Wei 2, Jun Yan 2, Yang Hu 2, Zhi-Hong Deng 1, Zheng Chen 2.
StoryFlow: Tracking the Evolution of Stories IEEE INFOVIS 2013 Shixia Liu, Senior Member, IEEE, Microsoft Research Asia Yingcai Wu, Member, IEEE, Microsoft.
The Table Lens: Merging Graphical and Symbolic Representations in an Interactive Focus+Context Visualization for Tabular Information Ramana Rao and Stuart.
Mingyang Zhu, Huaijiang Sun, Zhigang Deng Quaternion Space Sparse Decomposition for Motion Compression and Retrieval SCA 2012.
Graph Visualization and Beyond … Anne Denton, April 4, 2003 Including material from a paper by Ivan Herman, Guy Melançon, and M. Scott Marshall.
Region-Based Saliency Detection and Its Application in Object Recognition IEEE TRANSACTIONS ON CIRCUITS AND SYSTEM FOR VIDEO TECHNOLOGY, VOL. 24 NO. 5,
Learning from Positive and Unlabeled Examples Investigator: Bing Liu, Computer Science Prime Grant Support: National Science Foundation Problem Statement.
Hierarchical Matching with Side Information for Image Classification
Presented by Ho Wai Shing
Uncovering Clusters in Crowded Parallel Coordinates Visualizations Alimir Olivettr Artero, Maria Cristina Ferreiara de Oliveira, Haim levkowitz Information.
GENDER AND AGE RECOGNITION FOR VIDEO ANALYTICS SOLUTION PRESENTED BY: SUBHASH REDDY JOLAPURAM.
Jin Yan Embedded and Pervasive Computing Center
Learning Photographic Global Tonal Adjustment with a Database of Input / Output Image Pairs.
2D-LDA: A statistical linear discriminant analysis for image matrix
Contextual Snapshots: Enriched Visualization with Interactive Spatial Annotations Peter Mindek 1, Stefan Bruckner 2,1 and M. Eduard Gröller 1 1 Institute.
VisDB and Pixel Bar Charts Daniel A. Keim et al. ICS 280 Information Visualization Presented by Jeff Ridenour 4/16/02.
Shape2Pose: Human Centric Shape Analysis CMPT888 Vladimir G. Kim Siddhartha Chaudhuri Leonidas Guibas Thomas Funkhouser Stanford University Princeton University.
EE368 Final Project Spring 2003
Data Mining – Intro.
Urban Sensing Based on Human Mobility
Unsupervised Riemannian Clustering of Probability Density Functions
Nonparametric Semantic Segmentation
Cheng-Ming Huang, Wen-Hung Liao Department of Computer Science
CSE572, CBS598: Data Mining by H. Liu
Visual Analytics for Big Video Visualization
CSE572, CBS572: Data Mining by H. Liu
Background extraction with a coarse to fine approach
Paper Reading Dalong Du April.08, 2011.
Gradient Domain Salience-preserving Color-to-gray Conversion
CSE572: Data Mining by H. Liu
Presentation transcript:

Visual Abstraction and Exploration of Multi-class Scatterplots Haidong Chen, Wei Chen, Honghui Mei, Zhiqi Liu, Kun Zhou, Weifeng Chen, Wentao Gu, Kwan-Liu Ma

Motivations Multi-class scatterplots Aid comparison Make relative judgments

Motivations Multi-class scatterplots Aid comparison Make relative judgments Challenges Severe point overdraw

Motivations Multi-class scatterplots Aid comparison Make relative judgments Challenges Severe point overdraw Drawn order dependent

1. Existing Techniques Density Estimation Changing the Visual Channels Spatial Distortion Interactions

1. Existing Techniques a) Woodruff et al., 1998 d) Choo et al., 2013 c) Dan et al., 2010 b) Chen et al., 2009 a) Constant density visualizations of non-uniform distributions of data b) A novel interface for interactive exploration of DTI fibers c) Stacking graphic elements to avoid over-plotting d) Utopian: User-driven topic modeling based on interactive nonnegative matrix factorization Density Estimation Changing the Visual Channels Spatial Distortion Interactions

1. Existing Techniques b) Feng et al., 2010 a) Bachthaler et al., 2008 c) Mayorga et al., 2013 a) Continuous Scatterplot b) Matching Visual Saliency to Confidence in Plots of Uncertain Data c) Splatterplots: Overcoming Overdraw in Scatter Plots Density Estimation Changing the Visual Channels Spatial Distortion Interactions

1. Existing Techniques a) Keim et al., 1998 b) Keim et al., 2009 c) Janetzko et al., 2013 a) The Gridfit Algorithm: An Efficient and Effective Approach to Visualizing Large Amounts of Spatial Data b) Generalized scatter plot c) Enhancing scatter plots using ellipsoid pixel placement and shading Density Estimation Changing the Visual Channels Spatial Distortion Interactions

1. Existing Techniques a) Buring, et al., 2006 b) Yuan, et al Zooming Focus + Context Brushing Fisheye Dragging a) User interaction with scatterplots on small screens-a comparative evaluation of geometric-semantic zoom and fisheye distortion b) Dimension projection matrix/tree: Interactive subspace visual exploration and analysis of high dimensional data Density Estimation Changing the Visual Channels Spatial Distortion Interactions

2. Our Solution – Overview Point Density EstimationPoint Resampling Interactive ExplorationColor OptimizationShape Design

2. Our Solution – Visual Abstraction Point Density EstimationPoint Resampling Interactive ExplorationColor OptimizationShape Design

Point Resampling Background for multi-class blue noise sampling r Single-Class sampling r : minimum distance away from each other Multi-Class sampling r 00 Rx =Rx = r 01 r 02 Class 0 Class 1 Class 2 r 10 r 11 r 12 r 20 r 21 r 22 [Wei et al., 2010] r 01 r 10

Point Resampling Sampling spaces The input scatterplotContinuous sampling

Point Resampling Sampling spaces The input data pointsContinuous samplingDiscrete sampling

Point Resampling Zooming inconsistency Resampling once the view zoomed Zoom in Coarse zooming level Fine zooming level

Point Resampling Hierarchical sampling Coarse zooming level Zoom in: New points are added Fine zooming level Zoom out: New points are removed Coarse zooming level Samples generated at coarse zooming levels Samples generated at fine zooming levels The Core Idea: Using samples generated in previous coarse levels as anchor points to constrain sampling at a fine level The Core Idea: Using samples generated in previous coarse levels as anchor points to constrain sampling at a fine level

2. Our Solution – Abstractive Visualization Point Density EstimationPoint Resampling Interactive ExplorationColor OptimizationShape Design

Abstractive Visualization Point color optimization Color set in the CIELAB color space: Objective function: : the account of the divided blocks : the inter-class weight for the m-th block : the intra-class weight for the m-th block Color distance constraint:

Abstractive Visualization Point shape design Orientation of the shape:: the local linear regression coefficients Ellipse Dot-Line

2. Our Solution – Interactive Exploration Point Density EstimationPoint Resampling Interactive ExplorationColor OptimizationShape Design

The Exploration System Interaction tools: Highlighting Selection Painting Brushing Annotations Snapshot

3. Comparison Conventional multi-class scatterplot Splatterplots [ Mayorga et al., 2013 ] Splatterplots enhanced with additional noise Ours Point count in the selected region before abstraction: #(■) = 464 #(■) = 1796 #(■) = 2756 Point count in the selected region after abstraction: #(■) = 4 #(■) = 27 #(■) = 41 Density order before abstraction: #(■) < #(■) < #(■) Density order after abstraction: #(■) < #(■) < #(■) Relative density orders are preserved by our method

4. Evaluations Case studies The NBA Teams’ Shooting Positions Dataset in the Season 382,779 Mobile User Profile Dataset in a Month User study 32 synthetic and real datasets

4. Evaluations – Case Study I Conventional Multi-class Scatterplot Ours

4. Evaluations – Case Study II Conventional Multi-class ScatterplotOurs Dot-line Representation total-call-charge total-talk-time

4. Evaluations – User Study Profiles of the 32 datasets tested in our user study Compared schemes the conventional scatterplots (C) the color blending method (CB) the color weaving method (CW) our method (OURS) Participants 26 (19 males, 7 females) Tasks T1 Data classes identification T2 Relative densities recognition Datasets

4. Evaluations – User Study Results of scores for each task Tasks: T1 Data classes identification T2 Relative densities recognition

5. Conclusions and Future Work Conclusions A new visual abstraction approach for multi-class scatterplot

5. Conclusions and Future Work Conclusions A new visual abstraction approach for multi-class scatterplot A visual exploration system

5. Conclusions and Future Work Conclusions A new visual abstraction approach for multi-class scatterplot A visual exploration system Future Work A sophisticated point color optimization model

5. Conclusions and Future Work Conclusions A new visual abstraction approach for multi-class scatterplot A visual exploration system Future Work A sophisticated point color optimization model A non-uniform sampling scheme Conventional multi-class scatterplot Ours

Visual Abstraction and Exploration of Multi-class Scatterplots Different multi-class scatterplot schemes for the Person Activity dataset. (a) The conventional scatterplot. (b) The Splatterplots. (c) The Splatterplots with additional noise. (d) Our method.