Motion capture databases are now large, varied, and widely used. This course will cover database techniques useful for organizing, processing, and navigating such databases. Techniques will be presented in a general manner, and current and potential applications will be discussed.
Defining a good distance function is critical for many uses of a motion capture database. We will review the typical distance functions (Euclidean distances, Mahalanobis distance, and distance functions that employ time-warping techniques). We will also review techniques for feature extraction and dimensionality reduction (SVD/PCA, ICA), which can be used to construct more robust distance functions. For indexing, we will review fast database methods for finding similar motion capture sequences, such as R-trees, M-trees and their variants. We will also review time-series forecasting methods (e.g., ARIMA) for potential applications in feature extraction, segmentation and anomaly/discontinuity detection.
Applications that will be reviewed include dimensionality reduction for use in generating optimal motions and for visualization, fast nearest neighbor techniques for performance animation from a small set of markers, and time series forecasting techniques for use in behavior segmentation.
The target audience is researchers who want to get up to speed with the major tools for indexing and processing large motion-capture databases. Also, practitioners who want a concise, intuitive overview of the state of the art.
The prerequisite is B.Sc. in computer science or equivalent.
See http://www.cs.cmu.edu/~christos/TALKS/SIGGRAPH-07-tutorial/course0067.pdf for sample foils. The URL for this very page is http://www.cs.cmu.edu/~christos/TALKS/EG-07/tut-07-eg/syllabus.html
In SIGGRAPH 2007, August 2007, with ~40-50 attendees.
See http://www.cs.cmu.edu/~christos/TALKS/SIGGRAPH-07-tutorial/ for details and foils in pdf
Nancy Pollard is an Associate Professor in the Robotics Institute and the Computer Science Department at Carnegie Mellon University. She received her PhD in Electrical Engineering and Computer Science from the MIT Artificial Intelligence Laboratory in 1994. She was awarded an NSF CAREER grant in 2001 for research on ``Quantifying Humanlike Enveloping Grasps'', and the Okawa Research Grant in 2006 for her work on ``Dexterity and Natural Motion for Computer Graphics and Robotics'' Her primary research objective is to understand how to create natural motion for animated human characters and humanoid robots, with particular focus on grasping, manipulation and hands.
Jessica Hodgins joined the Robotics Institute and Computer Science Department at Carnegie Mellon University as a Associate Professor in fall of 2000. Prior to moving to CMU, she was an an Associate Professor and Assistant Dean in the College of Computing at Georgia Institute of Technology. She received her Ph.D. in Computer Science from Carnegie Mellon University in 1989. She has received a NSF Young Investigator Award, a Packard Fellowship, and a Sloan Fellowship. She was editor-in-chief of ACM Transactions on Graphics from 2000-2002 and Papers Chair for ACM SIGGRAPH 2003.
Christos Faloutsos is a Professor at Carnegie Mellon University. He has received the Presidential Young Investigator Award by the National Science Foundation (1989), the Research Contributions Award in ICDM 2006, eleven ``best paper'' awards, and several teaching awards. He has served as a member of the executive committee of SIGKDD; he has published over 160 refereed articles, 11 book chapters and one monograph. He holds five patents and he has given over 20 tutorials and 10 invited distinguished lectures. His research interests include data mining for streams and graphs, fractals, database performance, and indexing for multimedia and bio-informatics data.