Visualization to compare karate motion captures

SILVIA MABEL CASTRO; MARTÍN LEONARDO LARREA; DANA KARINA URRIBARRI; ENRICO PUPPO

Río Cuarto, Córdoba

Congreso; CACIC 2019; 2019

RedUNCI

Multi-dimensional time series from motion capture (MoCap) provide a rich source of data for human motion analysis, yet they are difficult to process and compare. We address MoCap data related to Karate katas, containing predefined sequences of movements, executed independently by several subjects with different timing and speed. We propose a combination of signal processing and data visualization techniques to analyze the misalignment between data from different subjects. We present a web app that implements this proposal, providing a visual comparisonof time series, based on Dynamic Time Warping.