動作辨識是動作擷取(motion capture) 處理中最受關注的題目之一。隨著近年來動作擷取裝置的普及，相關的應用也逐漸增加。目前常見的應用中，大多仍須仰賴人工定義連續動作中多個靜態姿勢(static pose)，或是特定關節位置，來代表完整的連續動作。此方法在動作變化較大的狀況下，若人工定義之靜態姿勢數量不足，或是動作擷取裝置之取樣率(sample rate) 過低，則容易產生誤判或是漏判之現象。而若以完整之動作資料代表本身以進行比對等相關應用，則會面臨計算過於耗時、動作資料取樣率不同及動作長度不同而無法逐一比對等問題。
　　為了解決上述問題，我們引入了關鍵姿勢自動萃取(key-pose extraction) 之概念，自動分析動作資料中進行動作轉換的時間點之動作資訊，作為該連續動作之特徵(feature)，用以代表完整之連續動作。如此能省去人工定義關鍵靜態姿勢之麻煩，又能僅以少量自動萃取出之動作資訊來代表完整動作，大大減少計算時間及避開目標與比對動作長短不同等問題。我們所提出之方法進行相似動作萃取之應用可大量減少人工定義關鍵姿勢所需時間及關鍵姿勢好壞之不確定性，以利後續用於門禁等安全系統中之身分辨識等應用。

Motion recognition is one of the most focused topic in motion captured data processing field. As the motion capture devices getting popular in recent years, the number of related applications also gets higher. Most of the applications rely on manually defined static poses or specific joint positions to represent the whole continuous motion sequence. This method suffers from insufficient key-poses in high-dynamic motions and low sampling rate of motion capture devices. On the other hand, if we use the whole motion sequence data for motion matching, the time consumption, variant motion lengths and distinct sampling rate problem would occur and decrease the quality of the result.
To conquer the difficuties above, we introduce the key-frame extraction method into this problem. We detect the motion transition positions automatically, and grab the nearby information as the feature to represent the whole motion sequence. Using this method, we can save not only manual works on defining static poses but also the computation time
due to only few local data is needed to represent the whole motion sequence, and avoid the problem of variant motion length. Our method can also be used for searching motions that is only similar on specific joints, to list all the variant motions based on user defined fixed joints. We also tested our method on handling variant skeleton parameters, to show that the key feature transform is robust enough to index between various skeletons.

[1] L. Kovar, M. Gleicher, and F. Pighin, ``Motion graphs,'' ACM Trans. Graph., vol. 21,
pp. 473--482, July 2002.
[2] J. Assa, Y. Caspi, and D. Cohen-Or, ``Action synopsis: pose selection and illustration,''
ACM Trans. Graph., vol. 24, pp. 667--676, July 2005.
[3] D. G. Lowe, ``Distinctive image features from scale-invariant keypoints,'' Int. J.
Comput. Vision, vol. 60, pp. 91--110, Nov. 2004.
[4] M.-W. Chao, C.-H. Lin, J. Assa, and T.-Y. Lee, ``Human motion retrieval from handdrawn
sketch,'' IEEE Transactions on Visualization and Computer Graphics, vol. 18,
pp. 729--740, May 2012.
[5] L. Kovar and M. Gleicher, ``Automated extraction and parameterization of motions
in large data sets,'' ACM Trans. Graph., vol. 23, pp. 559--568, Aug. 2004.
[6] K. Forbes and E. Fiume, ``An efficient search algorithm for motion data using
weighted pca,'' in Proceedings of the 2005 ACM SIGGRAPH/Eurographics symposium
on Computer animation, SCA '05, (New York, NY, USA), pp. 67--76, ACM,
2005.
[7] M. Muller, T. Roder, and M. Clausen, ``Efficient content-based retrieval of motion
capture data,'' ACM Trans. Graph., vol. 24, pp. 677--685, July 2005.
[8] B. Demuth, T. Roder, M. Muller, and B. Eberhardt, ``An Information Retrieval System
for Motion Capture Data,'' in In Proceedings of the 28th European Conference
on Information Retrieval (ECIR), vol. 3936, pp. 373--384, Mar. 2006.
[9] M. Muller and T. Roder, ``Motion templates for automatic classification and retrieval
of motion capture data,'' in Proceedings of the 2006 ACM SIGGRAPH/Eurographics
symposium on Computer animation, SCA '06, pp. 137--146, 2006.
[10] E. Keogh, T. Palpanas, V. B. Zordan, D. Gunopulos, and M. Cardle, ``Indexing large
human-motion databases,'' in Proceedings of the Thirtieth international conference
on Very large data bases - Volume 30, VLDB '04, pp. 780--791, 2004.
26
[11] O. C. JENKINS and M. J. Mataric, ``Performance-Derived Behavior Vocabularies:
Data-Driven Acquisition of Skills from Motion,'' vol. 1, pp. 237--288, 2004.
[12] E. Hsu, S. Gentry, and J. Popović, ``Example-based control of human motion,'' in
Proceedings of the 2004 ACM SIGGRAPH/Eurographics symposium on Computer
animation, SCA '04, pp. 69--77, 2004.
[13] G. Liu, J. Zhang, W. Wang, and L. McMillan, ``A system for analyzing and indexing
human-motion databases,'' in Proceedings of the 2005 ACM SIGMOD international
conference on Management of data, SIGMOD '05, pp. 924--926, 2005.
[14] G. Liu, J. Zhang, W. Wang, and L. McMillan, ``Human motion estimation from
a reduced marker set,'' in Proceedings of the 2006 symposium on Interactive 3D
graphics and games, I3D '06, (New York, NY, USA), pp. 35--42, ACM, 2006.
[15] J. Chai and J. K. Hodgins, ``Performance animation from low-dimensional control
signals,'' ACM Trans. Graph., vol. 24, pp. 686--696, July 2005.
[16] Y. Sakamoto, S. Kuriyama, and T. Kaneko, ``Motion map: image-based retrieval and
segmentation of motion data,'' in Proceedings of the 2004 ACM SIGGRAPH/Eurographics
symposium on Computer animation, SCA '04, (Aire-la-Ville, Switzerland,
Switzerland), pp. 259--266, Eurographics Association, 2004.
[17] L. Ren, Statistical analysis of natural human motion for animation. PhD thesis,
Pittsburgh, PA, USA, 2006.
[18] A. Safonova, J. K. Hodgins, and N. S. Pollard, ``Synthesizing physically realistic
human motion in low-dimensional, behaviorspecific spaces,'' ACM Transactions on
Graphics, pp. 514--521, 2004.
[19] O. Arikan, ``Compression of motion capture databases,'' ACM Trans. Graph., vol. 25,
pp. 890--897, July 2006.
[20] A. Shapiro, Y. Cao, and P. Faloutsos, ``Style components,'' in In Proc. of Graphics
interface, 2006.
[21] Y. Li, T. Wang, and H.-Y. Shum, ``Motion texture: a two-level statistical model for
character motion synthesis,'' ACM Trans. Graph., vol. 21, pp. 465--472, July 2002.
27
[22] K. Grochow, S. L. Martin, A. Hertzmann, and Z. Popović, ``Style-based inverse
kinematics,'' ACM Trans. Graph., vol. 23, pp. 522--531, Aug. 2004.
[23] A. Fod, M. Matari, and O. Jenkins, ``Automated derivation of primitives for movement
classification,'' Autonomous Robots, vol. 12, no. 1, pp. 39--54, 2002.
[24] J. Barbic, , J.-Y. Pan, C. Faloutsos, J. K. Hodgins, and N. Pollard, ``Segmenting
motion capture data into distinct behaviors,'' in In Proceedings of Graphics Interface
2004, pp. 185 -- 194, May 2004.
[25] O. Arikan, D. A. Forsyth, and J. F. O'Brien, ``Motion synthesis from annotations,''
ACM Trans. Graph., vol. 22, pp. 402--408, July 2003.
[26] J. Lee, J. Chai, P. Reitsma, J. K. Hodgins, and N. Pollard, ``Interactive control of
avatars animated with human motion data,'' ACM Transactions on Graphics (SIGGRAPH
2002), vol. 21, pp. 491 -- 500, July 2002.
[27] J. Wang and B. Bodenheimer, ``Computing the duration of motion transitions: an
empirical approach,'' in Proceedings of the 2004 ACM SIGGRAPH/Eurographics
symposium on Computer animation, SCA '04, pp. 335--344, 2004.
[28] O. Arikan and D. A. Forsyth, ``Interactive motion generation from examples,'' ACM
Trans. Graph., vol. 21, pp. 483--490, July 2002.
[29] L. Ikemoto, O. Arikan, and D. Forsyth, ``Quick transitions with cached multi-way
blends,'' in In ACM Symposium on Interactive 3D Graphics, pp. 145--151, 2007.
[30] L. Kovar, J. Schreiner, and M. Gleicher, ``Footskate cleanup for motion capture
editing,'' in Proceedings of the 2002 ACM SIGGRAPH/Eurographics symposium on
Computer animation, SCA '02, pp. 97--104, 2002.
[31] L. Ren, A. Patrick, A. A. Efros, J. K. Hodgins, and J. M. Rehg, ``A data-driven approach
to quantifying natural human motion,'' ACM Transactions on Graphics (SIGGRAPH
2005), vol. 24, pp. 1090--1097, Aug. 2005.