An efficient algorithm for compression of motion capture signal using multidimensional quadratic Bézier curve break-and-fit method

摘要: The emergence of applications to capture, process, store, and transmit motion capture (MoCap) signal has raise the interest in research community investigate devise better techniques for parameterization compression MoCap signal. In this work, we present a novel efficient method parametric representation skeletal animation. exploits temporal coherence using quadratic Bezier curve (QBC) fitting. treats rotational translation variations joint sequence frames as input points N-dimensional Euclidean space. are parameterized approximated QBC least square Break fit criterion is used minimize number segments required data. Precise control fitting accuracy achieved by user specified tolerance error limit. We compared performance proposed with principal component analysis wavelet transform based methods compression. leads smaller storage visual quality other methods. low degree ensures computationally algorithm, especially real-time applications.