Motion Representation Based on Important Turning Points Set and Its Application in Dance Training

Motion analysis is quite important in assistance of dance training, as it provides further analysis with basic information of the motion behavior. Since the scale of the motion time series are usually quite large, data compression becomes quite important. Thus, before the classification or similarity comparison, we need to analyze the data of each possible motion carefully and extract the important parts of the data to create the lower dimensional transformation of the motion data. In this paper, we propose an algorithm based on important turning points set which extracts data points that play significant roles in forming the shape of the motion data.