摘要: 运动捕获数据去噪旨在从含有噪声干扰的运动数据中恢复出能够较好表达原始数据特性的帧序列。针对人体运动捕获数据在较短时间段内的帧序列常常具有相同或相似的运动行为语义的特点,提出了一种分段式低秩逼近策略的运动捕获数据去噪方法。该方法首先将含有噪声的运动数据划分为多个连续子区间,接着利用不精确拉格朗日乘子法(IALM)对每个分段子区间的含噪声干扰数据批矩阵进行低秩矩阵逼近和稀疏噪声误差估计,达到分段数据去噪目的;最后利用时序特性组合去噪后的分段低秩逼近矩阵进行整体运动捕获数据去噪恢复。仿真实验结果表明,所提方法能够对含有任意拓扑结构的人体运动捕获数据进行去噪,达到了很好的效果,具有一定的通用性和实用性。
[1] Lee J,Shin S Y.General construction of time-domain filters for orientation data [J].IEEE Transactions on Visualization and Computer Graphics,2002,8(2):119-128 [2] Shin H J,Lee J,Shin S Y,et al.Computer puppetry:an importance-based approach [J].ACM Transactions on Graphics,2001,20(2):67-94 [3] Tak S,Ko H S.A physically-based motion retargeting filter [J].ACM Transactions on Graphics,2005,24(1):98-117 [4] Lou H,Chai J.Example-based human motion denoising [J].IEEE Transactions on Visualization and Computer Graphics,2010,16(5):870-879 [5] Park S I,Hodgins J K.Capturing and animating skin deformation in human motion [J].ACM Transactions on Graphics,2006,25(3):881-889 [6] Hsieh C C,Kuo P L.An impulsive noise reduction agent for rigidbody motion data using B-spline wavelets [J].Expert Systems with Applications,2008,34(3):1733-1741 [7] Tangkuampien T,Suter D.Human motion de-noising via greedy kernel principal component analysis filtering [C]∥Proc.18th IEEE International Conference on Pattern Recognition.2006:457-460 [8] Donoho,Candes,Tao T.Compressed sensing [J].IEEE Transa-ctions on Information Theory,2006,52(4):1289-1306 [9] Candes E J,Li X,Ma Y,et al.Robust principal component analysis? [J].Journal of the ACM,2011,58(3):1-31 [10] Lai R Y Q,Yuen P C,Lee K K W.Motion capture data completion and denoising by singular value thresholding [C]∥Proc.Eurographics Association.2011:45-48 [11] CMU Graphics Lab Motion Capture Database.http://mocap.cs.cmu.edu/ |
No related articles found! |
|