A Semantic Based Similarity Measure for Human Motion Data
In this paper, we measure the similarity of human motion data in the terms of distances between trajectories with a semantic method. In order to solve the problem of heavy computation cost, the semantic method that represents a trajectory as a set of 5-D vectors which contains the semantic information are proposed. Through experiments, the semantic method is proved to be efficient for cutting down the computation time and for two kinds of problems: overlapped trajectories with different directions and the trajectories with decoytrenches.
J. J. Zhao et al., "A Semantic Based Similarity Measure for Human Motion Data", Applied Mechanics and Materials, Vol. 235, pp. 384-388, 2012