@techreport{oai:ipsj.ixsq.nii.ac.jp:00141771,
 author = {蔡, 東生 and 董, 然 and 浅井, 信吉 and Dongshen, Cai and Dong, Ran and Nobushoshi, Asai},
 issue = {1},
 month = {May},
 note = {ヒルベルトーファン変換 （Hilbert-Huang Transform:HHT） は，経験的モード分解により，信号を複数の固有モード関数に分解し，ヒルベルト変換をかけ，時間周波数特性を分析する．時間周波数特性への鋭敏性は，フーリエ変換，ウエーブレット変換より遥かに鋭敏で，本報告では，多変量 HHT を用い，パヒューム，能楽，文楽などの動作を，ワルツ，ヒップホップ，サルサなどの踊りと比較する．, A new method for analyzing a set of multivariate nonlinear and non-stationary motion data has been developed. The key part of the method is the “empirical mode decomposition (EMD)” method with which any complicated multivariate motion data set can be decomposed into a finite multivariate “intrinsic mode functions (IMF)” that can Hilbert-transform later[Huang and Shen, 2005]. This motion decomposition method is adaptive and sensitive to noises. Since the motion decomposition is based on the local characteristic time scale of the multivariate joint data, it is applicable to nonlinear and non-stationary motion processes. Applying the Hilbert transform to the multivariate joint “intrinsic mode functions (IMF)” yield a set of instantaneous frequencies as functions of time that give imbedded structures of decomposed motions. The results are an energy-frequency-time distribution of joint motions, designated as the Hilbert spectrum. Using all three dancers' noisy motion capture data to decompose the motion, extremely complicate motions by the nonlinear and non-stationary effects, for example, dance turning motions, are clearly decomposed in the energy-frequency-time distribution.},
 title = {Perfumeのダンスはなぜ難しいのか？―多変量ヒルベルトーファン変換によるモーション解析},
 year = {2015}
}