Fast Subsequence Matching in Motion Capture Data

Motion capture data digitally represent human movements by sequences of body configurations in time. Subsequence matching in such spatio-temporal data is difficult as query-relevant motions can vary in lengths and occur arbitrarily in a very long motion. To deal with these problems, we propose a new subsequence matching approach which (1) partitions both short query and long data motion into fixed-size segments that overlap only partly, (2) uses an effective similarity measure to efficiently retrieve data segments that are the most similar to query segments, and (3) localizes the most query-relevant subsequences within extended and merged retrieved segments in a four-step postprocessing phase. The whole retrieval process is effective and fast in comparison with related work. A real-life 68-minute data motion can be searched in about 1 s with the average precision of \(87.98\%\) for 5-NN queries.

[1]  Reinhard Klein,et al.  Efficient unsupervised temporal segmentation of human motion , 2014, SCA '14.

[2]  Hans-Peter Seidel,et al.  Efficient and Robust Annotation of Motion Capture Data , 2009 .

[3]  Tido Röder,et al.  Documentation Mocap Database HDM05 , 2007 .

[4]  Michael Neff,et al.  Deep signatures for indexing and retrieval in large motion databases , 2015, MIG.

[5]  David Novak,et al.  Efficient Image Search with Neural Net Features , 2015, SISAP.

[6]  Dimitrios Gunopulos,et al.  Embedding-based subsequence matching in time-series databases , 2011, TODS.

[7]  Norman I. Badler,et al.  Efficient motion retrieval in large motion databases , 2013, I3D '13.

[8]  Pavel Zezula,et al.  Effective and efficient similarity searching in motion capture data , 2017, Multimedia Tools and Applications.

[9]  Christos Faloutsos,et al.  Fast subsequence matching in time-series databases , 1994, SIGMOD '94.

[10]  C.-C. Jay Kuo,et al.  Automatic Human Mocap Data Classification , 2014, IEEE Transactions on Multimedia.

[11]  Pavel Zezula,et al.  A Key-Pose Similarity Algorithm for Motion Data Retrieval , 2013, ACIVS.

[12]  Yuan Yan Tang,et al.  Efficient Human Motion Retrieval via Temporal Adjacent Bag of Words and Discriminative Neighborhood Preserving Dictionary Learning , 2017, IEEE Transactions on Human-Machine Systems.

[13]  Eamonn J. Keogh,et al.  Beyond one billion time series: indexing and mining very large time series collections with $$i$$SAX2+ , 2013, Knowledge and Information Systems.

[14]  Gang Wang,et al.  NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Pavel Zezula,et al.  Motion Images: An Effective Representation of Motion Capture Data for Similarity Search , 2015, SISAP.

[16]  Pavel Zezula,et al.  Similarity Searching in Long Sequences of Motion Capture Data , 2016, SISAP.

[17]  Yong Du,et al.  Hierarchical recurrent neural network for skeleton based action recognition , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Thomas Seidl,et al.  Efficient Query Processing in 3D Motion Capture Databases via Lower Bound Approximation of the Gesture Matching Distance , 2015, 2015 IEEE International Symposium on Multimedia (ISM).