“3D human motion compression using wavelet decomposition” by Lee and Lasenby

  • ©Chao-Hua Lee and Joan Lasenby

  • ©Chao-Hua Lee and Joan Lasenby

Conference:


Type(s):


Title:

    3D human motion compression using wavelet decomposition

Presenter(s)/Author(s):



Abstract:


    Motion capture is the richest source of data on human motion for computer animation. The data always contains a large number of frames which translates to large size files, and makes post-processing difficult. While there has been significant research into compression of mesh animation [Alexa and Muller 2000], little  ̈  work on compression of mocap data exists.  We propose a novel framework based on multiresolution wavelet decomposition for compressing motion capture data. Given a global error constraint, our system finds the optimal combination of the decomposition levels for the joints that minimize the number of frames stored. Our approach can achieve at least 30 to 1 compression ratio in the reconstruction without noticeable perceptual artifacts; eg negligible foot skate is observed.

References:


    1. Alexa, M., and Müller, W. 2000. Representing animations by principal components. Comput. Graph. Forum 19, 3.
    2. Bellman, R. 1957. Dynamic programming. Princeton Univesity Press.
    3. Millat, S. 1989. A theory for multiresolution signal decomposition: The wavelet representation. IEEE Trans. Pattern Anal. Mach. Intell. 11, 7, 674–693.


ACM Digital Library Publication:



Overview Page: