Motion frequency data analysis for sports skill
Toshiyuki Maeda, Masumi Yajima · 2017 Computing Conference · 2017
This paper addresses motion frequency data analysis for sports skill, focused on volleyball attack skill. We attempt to prove a hypothesis that expert skills have lower frequency motions, rather than the similarity of novice skills as human gesture control. To this end, we use volleyball to attack the time series of moving images for experimentation and exercise techniques to analyze the movement frequency. In our study, using high-speed cameras to record motion image data to analyze volleyball matches, we do not use physical information like body skeleton models. The time series data is measured from moving image data with four marker points and analyzed by fast Fourier transform (FFT) as well as Mel-frequency cepstral coefficients (MFCC), using clustering methods. The experimental results show that novice data has higher frequency data, which means that novice moves with high frequency motion, which may help us to assume that expert players use low frequency motion.