An efficient search algorithm for motion data using weighted PCA

Good motion data is costly to create. Such an expense often makes the reuse of motion data through transformation and retargetting a more attractive option than creating new motion from scratch. Reuse requires the ability to search automatically and efficiently a growing corpus of motion data, which remains a difficult open problem. We present a method for quickly searching long, unsegmented motion clips for subregions that most closely match a short query clip. Our search algorithm is based on a weighted PCA-based pose representation that allows for flexible and efficient pose-to-pose distance calculations. We present our pose representation and the details of the search algorithm. We evaluate the performance of a prototype search application using both synthetic and captured motion data. Using these results, we propose ways to improve the application's performance. The results inform a discussion of the algorithm's good scalability characteristics.

[1]  David Salesin,et al.  Fast multiresolution image querying , 1995, SIGGRAPH.

[2]  Lance Williams,et al.  Motion signal processing , 1995, SIGGRAPH.

[3]  Eamonn J. Keogh,et al.  Iterative Deepening Dynamic Time Warping for Time Series , 2002, SDM.

[4]  Jernej Barbic,et al.  Segmenting Motion Capture Data into Distinct Behaviors , 2004, Graphics Interface.

[5]  Dimitrios Gunopulos,et al.  Indexing Large Human-Motion Databases , 2004, VLDB.

[6]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[7]  Eamonn J. Keogh,et al.  Scaling up Dynamic Time Warping to Massive Dataset , 1999, PKDD.

[8]  Michael Patrick Johnson,et al.  Exploiting quaternions to support expressive interactive character motion , 2003 .

[9]  Michael Gleicher,et al.  Automated extraction and parameterization of motions in large data sets , 2004, SIGGRAPH 2004.

[10]  Lucas Kovar,et al.  Flexible automatic motion blending with registration curves , 2003, SCA '03.

[11]  Harry Shum,et al.  Motion texture: a two-level statistical model for character motion synthesis , 2002, ACM Trans. Graph..

[12]  Daniel Thalmann,et al.  PCA-based walking engine using motion capture data , 2004, Proceedings Computer Graphics International, 2004..

[13]  Dimitrios Gunopulos,et al.  Discovering similar multidimensional trajectories , 2002, Proceedings 18th International Conference on Data Engineering.

[14]  Michael Neff,et al.  Methods for exploring expressive stance , 2004, SCA '04.

[15]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[16]  Daniel Thalmann,et al.  PCA-based walking engine using motion capture data , 2004 .

[17]  Alex Pentland,et al.  Invariant features for 3-D gesture recognition , 1996, Proceedings of the Second International Conference on Automatic Face and Gesture Recognition.

[18]  Sam T. Roweis,et al.  EM Algorithms for PCA and SPCA , 1997, NIPS.

[19]  David A. Forsyth,et al.  Motion synthesis from annotations , 2003, ACM Trans. Graph..

[20]  F. Sebastian Grassia,et al.  Practical Parameterization of Rotations Using the Exponential Map , 1998, J. Graphics, GPU, & Game Tools.

[21]  Lucas Kovar,et al.  Motion graphs , 2002, SIGGRAPH '08.

[22]  François Bérard,et al.  Bare-hand human-computer interaction , 2001, PUI '01.

[23]  Daniel Thalmann,et al.  A Coherent Locomotion Engine Extrapolating Beyond Experimental Data , 2004 .

[24]  Sunil Arya,et al.  An optimal algorithm for approximate nearest neighbor searching fixed dimensions , 1998, JACM.

[25]  D. Skočaj.  Weighted Incremental Subspace Learning , 2002 .

[26]  Dimitrios Gunopulos,et al.  Fast Motion Capture Matching with Replicated Motion Editing , 2003 .

[27]  Fernando Pereira,et al.  MPEG-7 the generic multimedia content description standard, part 1 - Multimedia, IEEE , 2001 .

[28]  Lucas Kovar,et al.  Automated extraction and parameterization of motions in large data sets , 2004, ACM Trans. Graph..