운동학적 접근 방법을 사용한 복잡한 인간 동작 질의 시스템 = A Kinematic Approach to Answering Similarity Queries on Complex Human Motion Data

다국어 초록 (Multilingual Abstract)

Recently there has arisen concern in both the database community and the graphics society about data retrieval from large motion databases because the high dimensionality of motion data implies high costs. In this circumstance, finding an effective distance measure and an efficient query processing method for such data is a challenging problem. This paper presents an elaborate motion query processing system, SMoFinder (Similar Motion Finder), which incorporates a novel kinematic distance measure and an efficient indexing strategy via adaptive frame segmentation. To this end, we regard human motions as multi-linkage kinematics and propose the weighted Minkowski distance metric. For efficient indexing, we devise a new adaptive segmentation method that chooses representative frames among similar frames and stores chosen frames instead of all frames. For efficient search, we propose a new search method that processes k-nearest neighbors queries over only representative frames. Our experimental results show that the size of motion databases is reduced greatly (x1/25) but the search capability of SMoFinder is equal to or superior to that of other systems.

참고문헌 (Reference)

1 H. T. Shen, "Towards effective indexing for very large video sequence database" ACM 730-741, 2005

2 Hyungsoo Jung, "SMoFinder: A System for Querying Complex Human Motions Using a Kinematic Approach" 640-643, 2008

3 L. Chen, "Robust and fast similarity search for moving object trajectories" ACM 491-502, 2005

4 A. Guttman, "R-trees: a dynamic index structure for spatial searching" 47-57, 1984

5 Y. Chen, "Querying complex spatio-temporal sequences in human motion databases" 90-99, 2008

6 P. Ciaccia, "M-tree: An efficient access method for similarity search in metric spaces" 426-435, 1997

7 E. Keogh, "Locally adaptive dimensionality reduction for indexing large time series databases" ACM 151-162, 2001

8 M. Vlachos, "Indexing multi-dimensional time-series with support for multiple distance measures" ACM 216-225, 2003

9 E. Keogh, "Indexing large human-motion databases" VLDB Endowment 780-791, 2004

10 S. Salvador, "Fastdtw: Toward accurate dynamic time warping in linear time and space" 2004

11 C. Faloutsos, "Fast subsequence matching in time-series databases" ACM 419-429, 1994

12 K. pong Chan, "Efficient time series matching by wavelets" 126-133, 1999

13 R. Agrawal, "Efficient similarity search in sequence databases" Springer-Verlag 69-84, 1993

14 M. Müller, "Efficient content-based retrieval of motion capture data" 24 (24): 677-685, 2005

15 P. N. Yianilos, "Data structures and algorithms for nearest neighbor search in general metric spaces" Society for Industrial and Applied Mathematics 311-321, 1993

16 "CMU graphics lab motion capture database"

17 L. Kovar, "Automated extraction and parameterization of motions in large data sets" 23 (23): 559-568, 2004

18 Acclaim, "ASF/AMC file specifications"