Three-dimensional human body movement data dividing method

Abstract

The invention discloses a partition method on human motion 3D data. First, use manifold analysis method ISOMAP to map the original human sequence motion data to the low dimensional manifold space; secondly, utilize a heuristic method to detect rough partition points of different types of movement in the motion data sequence, then long sequence of human motion data are partitioned to several segments; last, for every two segments concatenated in time in the low dimensional manifold space use K-average clustering algorithm to compute the accurate partition point.

Description

technical field [0001] The invention relates to the field of computer three-dimensional animation technology and multimedia data processing, in particular to a method for segmenting three-dimensional human motion data. Background technique [0002] In recent years, with the widespread use of motion capture devices, a large amount of realistic 3D human motion data has been generated, and these data are widely used in computer games, animation generation, and medical simulation and other fields. Since the human body motion sequence obtained when the optical capture system is used to capture in-body motion is relatively long, and often contains several continuous but different types of motion data (such as walking, running, jumping, etc.), in order to facilitate a large number of 3D human motion data It is very useful for automatic segmentation of different types of motion data contained in long-sequence human motion data. [0003] Data segmentation is a technique widely used ...

AI Technical Summary

Problems solved by technology

The model-based motion data segmentation method needs to be based on a large amount of training data, and the segmentation results are greatly affected by the training data, so it is difficult to be applied in practice

[0004] The method in the first issue of "Journal of Autonomous Robotics" in 2002 (Autonomous Robots, 2002, 12 (1): 39-54) uses the method of detecting the zero-crossing point of the joint angle data to segment the human arm motion data. This method is very simple to implement. But the accuracy of the segmentation results is not high

The method published in the Proceedings of Graphics Interface Conference in 2005 (Proceedings of Graphics Interface, 2004, 185-194) believes that different types of human motion data should have different internal dimensions, so methods such as PCA, PPCA, and GMM are used to analyze long-sequence human body Dimensional analysis of motion data, and then achieve the purpose of data segmentation, this method can achieve high segmentation accuracy, and does not require training data support, but the construction of PCA, PPCA subspace and the establishment of statistical models such as GMM take too much time , so the segmentation efficiency is not high

Images