Selected Publications (Full List Here)
-
Analysis of Stock Market Bubble Burst with Level-Of-Detail Visualization
Sruthi Iyer, Joseph Rakestraw, Yong Cao, Steven Sheetz and Raman KumarInternational Working Conference on Advanced Visual Interfaces AVI 2014 (In Submission)
In this paper, we present a circle packed layout along with two levels of detail for the sectors and firms, to visualize and analyze market capitalization values which enables regulators to reach educated conclusions regarding market trends and conditions.
pdf (Available later) video
Hardware-Aware Parallel Algorithms for Direct Volume Rendering on GPU
Junpeng Wang, Fei Yang and Yong CaoEuroVIS 2014 (In Submission)
In this paper, we analyze the volume ray casting algorithm on three important hardware related performance issues of GPUs: cache performance, branching efficiency and load balancing. To address these issues, we present a new sampling strategy in a parallel algorithm, called warp marching. Our algorithm displays a novel computation-to-core mapping method which samples the volume data in a cache-friendly manner. In addition, we introduce a double-buffer approach and leverage special GPU operations, such as warp shuffling, to improve the efficiency of parallel execution.
pdf (Available later) video
-
StockBubbler: A Tool for Stock Market Bubble Visualization
Sruthi Iyer, Joseph Rakestraw, Yong Cao, Steven Sheetz and Raman KumarGraphics Interface 2014 (In Submission)
StockBubbler is an interactive visualization tool for observing stock market data and predicting the onset of a stock market bubble burst. StockBubbler can be used by financial regulators to study market capitalization data through the circle packed design of StockBubbler along with the geometric mean return and percentage change in performance values to detect a potential bubble in the system.
pdf (Available later) video
-
Simulating and Animating Social Dynamics: Embedding Small Pedestrian Groups in Crowds
Seung In Park, Francis Quek and Yong CaoComputer Animation and Social Agents 2013 &
Computer Animation and Virtual Worlds
We present a crowd model informed by common ground theory to accommodate high-level socially aware behavioral realism of characters in crowd simulations. In our approach, group members maintain group cohesiveness by communicating and adapting their behaviors to each other. The resulting character behaviors in animations form a consequential chain interpreted as a coherent story by observers. We demonstrate that our model produces more believable animations from the viewpoint of human observers through a series of user studies.
-
Integrating Occlusion Culling with Parallel LOD for Rendering Complex 3D Environments on GPU
Chao Peng and Yong CaoI3D 2013 Poster Paper
We present a novel rendering approach that seamlessly integrates parallel LOD algorithm, parallel occlusion culling and Out-of-Core method in a unified scheme towards GPU architectures. The result shows the parallel occlusion culling significantly reduces the required complexity of the 3D model and increases the rendering performance.
-
Multi-GPU Load Balancing for In-Situ Simulation and Visualization
Yong Cao and Robert HaganPDPTA 2011 & Journal of SuperComputing 2013 (In Major Revision)
The aim of this paper is to research on the memory management and scheduling issues in the multi-GPU environment, in order to balance the workload between this two-stage pipeline execution. We first propose a data-driven load balancing scheme which takes into account of some important performance factors for scientific simulation and rendering, such as the number of iterations for the simulation and the rendering resolution. As an improvement to this scheduling method, we also introduce a dynamic load balancing approach that can automatically adjust the workload changes at runtime to achieve better load balancing results. This approach is based on an idea to analytically approximate the execution time difference between the simulation and the rendering by using fullness of the synchronization data buffer.
-
Modeling Social Groups in Crowds Using Common Ground Theory
Seung In Park, Francis Quek and Yong CaoWinter Simulation Conference 2012
In this paper we advance a computational model informed by Common Ground (CG) Theory that both inherits the social realism provided by the CG model and is computationally tractable for a large number of groups and individuals. The task of navigation in a group is viewed as performing a joint activity among agents, which requires effective coordination among group members. Our model includes both macro and micro coordination, addressing the joint plans, and the actions for coordination respectively. These coordination activities and plans inform the high-level route and walking strategies of the agents.
-
Load Balanced Parallel GPU Out-of-Core for Continuous LOD Model Visualization
Chao Peng, Mi Peng and Yong CaoUltraVis 2012
IThis paper explores a device-level parallel design that distributes the workloads for both GPU out-of-core and LOD processing in a multi-GPU multi-display system. Our multi-GPU out-of-core takes advantages of a load-balancing method and seamlessly integrates with the parallel LOD algorithm. By usingframe-to-frame coherence, the overhead of data transferring is significantly reduced on each GPU.
-
Modeling Agent Social Joint Actions via Micro and Macro Coordination Strategies
Seung In Park, Francis Quek and Yong CaoIntelligent Agent Technology (IAT) 2012
Past research on multi-agent simulation with cooperative reinforcement learning (RL) focuses on developing sharing strategies that are adopted and used by all agents in the environment. In this paper, we target situations where this assumption of a single sharing strategy that is employed by all agents is not valid. We seek to address how agents with no predetermined sharing partners can exploit groups of cooperatively learning agents to improve learning performance when compared to Independent learning. Specifically, we propose 3 intra-agent methods that do not assume a reciprocating sharing relationship and leverage the pre-existing agent interface associated with Q-Learning to expedite learning.
-
Non-Reciprocating Sharing Methods in Cooperative Q-Learning Environments
Bryan Cunningham and Yong CaoIntelligent Agent Technology (IAT) 2012
Past research on multi-agent simulation with cooperative reinforcement learning (RL) focuses on developing sharing strategies that are adopted and used by all agents in the environment. In this paper, we target situations where this assumption of a single sharing strategy that is employed by all agents is not valid. We seek to address how agents with no predetermined sharing partners can exploit groups of cooperatively learning agents to improve learning performance when compared to Independent learning. Specifically, we propose 3 intra-agent methods that do not assume a reciprocating sharing relationship and leverage the pre-existing agent interface associated with Q-Learning to expedite learning.
-
A Versatile Optical Model for Hybrid
Fei Yang, Qingde Li, Dehui Xiang, Yong Cao and Jie Tian
Rendering of Volume DataIEEE TVCG 2012
In this paper, we propose a generalized optical model that combines particle elements and surface elements together to characterize both the behavior of individual particles and the collective effect of particles. The framework based on a new model provides a more powerful and flexible tool for hybrid rendering of isosurfaces and transparent clouds of particles in a single scene. It also provides a more rational basis for shading, so the problem of normal-based shading in homogeneous regions encountered in conventional volume rendering can be easily avoided.
-
A Crowd Modeling Framework for Socially Plausible Animation Behaviors
Seung In Park, Chao Peng, Francis Quek and Yong CaoMotion In Games MIG 2012
This paper presents a framework for crowd modeling that produces socially plausible animation behaviors. Our high-level behavioral model is able to produce appropriate animated behavior that includes synchronized body-orientation and gesture of individual actors within the simulation.
-
A GPU-based Approach for Massive Model Rendering with Frame-to-Frame Coherence
Chao Peng and Yong CaoEurographics 2012 & Computer Graphics Forum
We present a GPU-based approach which enables interactive rendering of large 3D models with hundreds of millions of triangles. Our work contributes to the massive rendering research in two ways. First, we present a simple and efficient
mesh simplification algorithm towards GPU architecture. Second, we propose a novel GPU out-of-core approach that adopts a frame-to-frame coherence scheme in order to minimize the high communication cost between CPU and GPU. -
Dynamic Analysis of Large Datasets with Animated and Correlated Views
Yong Cao, Reese Moore, Peng Mi, Alex Endert, Chris North, Randy MarchanyVAST 2012 Mini Challenge Award
In this paper, we introduce a GPU-accelerated visual analytics tool, AVIST. By adopting the in-situ visualization architecture on the GPUs, AVIST supports real-time data analysis and visualization of massive scale datasets, such as VAST 2012 Challenge dataset. The design objective of the tool is to identify temporal patterns from large and complex data. To achieve this goal, we introduce three unique features: automatic animation, disjunctive data filters, and time-synced visualization of multiple datasets.
-
Modeling Small Group Behaviors in Large Crowd Simulation
Seung In Park, Yong Cao and Francis QuekI3D 2012 Poster Paper
This paper presents a framework to simulate many number of small groups. To achieve this, we introduce a local interaction field(LIF) which embeds time-space information of surrounding environment of each group into the potential fields of a continuum dynamics simulation.
-
Levels of Realism for Cooperative
Bryan Cunningham and Yong Cao
Multi-agent Reinforcement LearningSwam Intelligence Conference 2012
Training agents in a virtual crowd to achieve a task can be accomplished by allowing the agents to learn by trial-and-error and by sharing information with other agents. Since sharing enables agents to potentially reach optimal behavior more quickly, what type of sharing is best to use to achieve the quickest learning times? This paper categorizes sharing into three categories: realistic, unrealistic, and no sharing. Realistic sharing is defined as sharing that takes place amongst agents within close proximity and unrealistic sharing allows agents to share regardless of physical location.
-
A Real-time System of Crowd Rendering:
Chao Peng, Seung In Park and Yong Cao
Parallel LOD and Texture-Preserving Approach
on GPUMotion In Games MIG 2011
We present a real-time crowd rendering system on GPUs with a special focus on how to preserve texture appearance in progressive LOD-based mesh simpli cation algorithms. Our results show that the proposed parallel LOD approach can get up to 5.33 times of speedup compared with the standard pseudo-instancing approach.
-
Large Scale Crowd Simulation Using A Hybrid Agent Model
Seung In Park, Yong Cao and Francis QuekMotion In Games MIG 2011
We present a hybrid model for large-scale crowd simulation by augmenting continuum dynamics under an agent based perspective. Our model supports both group behaviors and complex behaviors driven by individual decisions of agents. Our simulation system is implementedon the parallel architecture of graphics processing units (GPUs), and scales very well with respect to the size of the virtual environment.
-
Many-core Architecture Oriented Parallel
Yong Cao
Algorithm Design for Computer AnimationMition in Games MIG 2011 (Invited Paper)
Many-core architecture has become an emerging and widely adopted platform for parallel computing. Computer animation researches can harness this advance in high performance computing with better understanding of the architecture and careful consideration of several important parallel algorithm design issues, such as computation-to-core mapping, load balancing and algorithm design paradigms. In this paper, we use a set of algorithms in computer animation as the examples to illustrate these issues, and provide possible solutions for handling them. We have shown in our previous research projects that the proposed solutions can greatly enhance the performance of the parallel algorithms.
-
A Real-time Algorithm for Search-based Motion
Chao Peng & Yong Cao
SynthesisCGVR 2011
The performance of the most of existing motion search algorithms do not scale well with respect to the size of the input constraints. We present a novel search algorithm, Single Stepping, that can provide a real-time solution when searching a large motion database. The algorithm starts with a greedy search result and can enumerate all possible optimal solutions in linear time with respect to the input constraints. A smoothness metric is also introduced to find the best resulting motion from theseoptimal solutions.
-
Adaptive InteractiveMulti-Resolution Computational Steering for Complex Engineering Systems
Kresimir Matkovic and Denis Gracanin and Mario Jelovic and Yong CaoEuroVA 2011
We describe an approach that, in aniterative manner, allows a domain expert to interactively select data points (design of experiments), approximate the values in a continuous region of the simulation space (regression) and automatically find the “best” points in that continuous region based on the specified constraints and objectives (optimization), using the regression and aggregated data. Once the objectives are found, data points in the neighborhood of the objective are generated by the simulation tool thus providing denser coverage of the regions of interest.
-
A Novel Computation-to-core Mapping Scheme for Robust Facet Image Modeling on GPUs
Yong Cao, Seung-In Park and Layne T. WatsonPDPTA 2010 & Journal of Real-Time Image Processing 2012
This paper shows how to optimize the computational parallelism in robust facet image modeling to GPU architecture, using finegrained block level parallelism achieved by assigning more GPU cores/threads to process one pixel, rather than pixel level parallelism.
-
A Motion Graph Approach for Interactive 3D Animation using Low-cost Sensors
Yong Cao and Mithilesh KumarCGVR 2010
This paper presents a motion graph based framework to produce high quality motion sequences in real-time using a set of inertial sensor based controllers. The user’s action generates signals from the controllers that provide constraints to select appropriate sequence of motions from a structured database of human motions, namely motion graph.
-
GPU Assisted Real-Time Isosurface Volume Rendering Using Depth Based Coherence and Variance Bricking
Colin Braley, Robert Hagan, Yong Cao and Denis GracaninCGVR 2010 & SIGGRAPH Asia 2009 Poster
We propose a novel GPU based dataset traversal technique that uses a prediction buffer to reduce the traversal time during dataset rotation. We use a highly parallelized ray-casting algorithm and the proposed traversal technique to double the rendering speed.
-
Towards Chip-on-Chip Neuroscience: Fast Mining of Neuronal Spike Streams Using Graphics Hardware
Yong Cao, Debprakash Patnaik, Sean Ponce, Jeremy Archuleta, Patrick Butler, Wu-chun Feng and Naren RamakrishnanComputing Frontier 2010 and GPU Computing Gem (Chapter 15)
Mining neuronal spike streams in computational neuroscience research is critical to understand the firing patterns of neurons and gain insight into the under- lying cellular activity. We present a solution that uses a massively parallel graphics processing unit (GPU) to mine the stream of spikes. Our contributions include new computation-to-core mapping schemes and novel strategies to map finite state machine-based counting algorithms onto the GPU.
-
Interacting with Stories
Ashley Robinson, Chao Peng, Francis Quek and Yong CaoWOCCI '09: The 2nd Workshop on Child, Computer and Interaction
In todays media-saturated world, students are consuming media both actively and passively. To facilitate active interaction with media, we address a specific kind of audio-visual media interaction in which we call a hyper-drama. These hyper-drama interactions include a token on a horizontal display versus mouse on a desktop display for story navigation, desktop display versus tablet display for scene viewing, and virtual buttons versus speech for character interaction and decision making within the hyper-drama.
-
Accelerator-Oriented Algorithm Transformation for Temporal Data Mining
Debprakash Patnaik, Sean P. Ponce, Yong Cao and Naren RamakrishnanIFIP International Conference on Network and Parallel Computing Workshops 2009
In todays media-saturated world, students are consuming media both actively and passively. To facilitate active interaction with media, we address a specific kind of audio-visual media interaction in which we call a hyper-drama. These hyper-drama interactions include a token on a horizontal display versus mouse on a desktop display for story navigation, desktop display versus tablet display for scene viewing, and virtual buttons versus speech for character interaction and decision making within the hyper-drama.
-
GPU Accelerated Fuzzy Connected Image Segmentation by using CUDA
Ying Zhuge, Yong Cao, Jayaram K. Udupa and Robert W. MillerThe 31st Annual International Conference of the IEEE Engineering in Medical and Biology Society & Medical Physics 2011
In this paper, we present a parallel fuzzy connected image segmentation algorithm implementation on Nvidia’s Compute Unified Device Architecture (CUDA) platform for segmenting large medical image data sets. Our experiments based on three data sets of small, medium, and large data size demonstrate the efficiency of the parallel algorithm.
-
Multi-Dimensional Characterization of Temporal Data Mining on Graphics Processors
Jeremy Archuleta, Yong Cao, Wu-chun Feng and Tom ScoglandIPDPS 2009
We present a characterization of a MapReduce-based data-mining application on a general-purpose GPU (GPGPU). Using neuroscience as the application vehicle, the results of our multidimensional performance evaluation show that an “one-size-fits-all” approach maps poorly across different GPGPU cards. Rather, a high-performance implementation on the GPGPU should factor in the 1) problem size, 2) type of GPU, 3) type of algorithm, and 4) data-access method when determining the type and level of parallelism.
-
Full Body Tracking Using An Agent-Based Architecture
Bing Fang, Liguang Xie, Pak-Kiu Chung, Yong Cao and Francis QuekIEEE 37th Applied Imagery Pattern Recognition Workshop
The paper presents an agent-based full body tracking and 3D animation system to generate motion data using stereo calibrated cameras. The power of this approach is the flexibility by which domain information may be encoded within each agent to produce an overall tracking solution.
-
Low-Cost, High-Speed Computer Vision Using NVIDIA's CUDA Architecture
Seung In Park, Sean Ponce, Jing Huang, Yong Cao and Francis QuekIEEE 37th Applied Imagery Pattern Recognition Workshop
In this paper, we introduce real time image processing techniques using modern programmable Graphic Processing Units (GPU). We demonstrate the efficiency of our approach by a parallelization and optimization of Canny’s edge detection algorithm, and applying it to a computation and data-intensive video motion tracking algorithm known as “Vector Coherence Mapping” (VCM).
-
A Nonlinear Manifold Learning Framework for Real-time Motion Estimation using Low-cost Sensors
Liguang Xie, Bing Fang, Yong Cao and Francis QuekIEEE 37th Applied Imagery Pattern Recognition Workshop
We propose a real-time motion synthesis framework to control the animation of 3D avatar in real-time. We apply a non-linear manifold learning method to establish a high dimensional motion model which learned from a large motion capture database. Then, by taking 3D accelerometer sensor signal as input, we are able to synthesize high-quality motion from the motion model we learned from the previous step.
-
GPU-accelerated computation for robust motion tracking using the CUDA framework
Jing Huang, Sean Ponce, Seung In Park, Yong Cao and Francis QuekThe 5th IET Visual Information Engineering 2008 Conference
In this paper, we discuss an implementation of a graphics hardware acceleration of the Vector Coherence Mapping vision processing algorithm. Using this algorithm as our test case, we discuss our optimization strategy for various vision processing operations using NVIDIA’s new CUDA programming framework.
-
Data-Driven Motion Estimation with Low-Cost Sensors
Liguang Xie, Mithilesh Kumar, Yong Cao, Denis Gracanin and Francis QuekThe 5th IET Visual Information Engineering 2008 Conference
We propose a motion estimation framework that utilizes a small set of low-cost, 3D acceleration sensors. A data-driven approach is used to synthesize realistic human motion comparable in quality to the motion captured by the professional motion capture systems.
-
Trauma Room Simulation
Yong Cao, Doug Bowman and Francis QuekWe worked with Carilion Memorial Hospital at Roanoke, VA and developed this system to train residence doctors how to handle emergency situation in Trauma room. The system includes animation, audio, simulation and camera components. The simulation is data-driven, so that the senior doctor can author different training senarios with the tools we provide. The system has been deployed to the hospital for training sessions. The future plan has been decided for more medical and educational training.
-
Style Components
Ari Shapiro and Yong Cao and Petros FaloutsosGI '06: Graphics Interface 2006
We propose a novel method for interactive editing of motion data based on motion decomposition. Our method employs Independent Component Analysis (ICA) to separate motion data into visually meaningful components called style components. The user then interactively identifies suitable style components and manipulates them based on a proposed set of operations.
-
Expressive Speech-Driven Facial Animation
Yong Cao, Wen Tien, Petros Faloutsos, Fred PighinACM Transaction on Graphics 2005
Issue 4 Volume 24We address expressive facial synthesis problem by using a machine learning approach that relies on a database of speech related high- delity facial motions. From this training set, we derive a generative model of expressive facial motion that incorporates emotion control while maintaining accurate lip-synching.
pdf video video_demo (a synthesized monologue)
-
Interactive Motion Decomposition
Ari Shapiro and Yong Cao and Petros FaloutsosACM Siggraph 2004 Sketch
We address expressive facial synthesis problem by using a machine learning approach that relies on a database of speech related high-fidelity facial motions. From this training set, we derive a generative model of expressive facial motion that incorporates emotion control while maintaining accurate lip-synching.
-
Real-time speech motion synthesis from recorded motions
Yong Cao, Petros Faloutsos, Eddie Kohler and Fred PighinSCA 2004: ACM SIGGRAPH/Eurographics symposium on Computer animation
Motion synthesis algorithms that are based on a graph of motions are notorious for their exponential complexity. In this paper, we present a greedy graph search algorithm that yields vastly superior performance and allows real-time motion synthesis from a large database of motions.
-
Unsupervised learning for speech motion editing
Yong Cao, Petros Faloutsos and Fred PighinSCA 2003: ACM SIGGRAPH/Eurographics symposium on Computer animation
We present a new method for editing speech related facial motions. Our method uses an unsupervised learning technique, Independent Component Analysis (ICA), to extract a set of meaningful parameters without any annotation of the data. With ICA, we are able to solve a blind source separation problem and describe the original data as a linear combination of two sources. One source captures content (speech) and the other captures style (emotion). By manipulating the independent components we can edit the motions in intuitive ways.