CONICET | Buscador de Institutos y Recursos Humanos

Motion capture (MoCap) data as time series provide a rich source of input for human movement analysis; however, their multidimensional nature makes them difficult to process and compare. In this paper, we propose a visual analysis technique that allows the comparison of MoCap data obtained from karate katas. These consist of a series of predefined movements that are executed independently by several subjects at different times and speeds. For the comparative analysis, the proposed solution presents a visual comparison of the misalignment between a set of time series, based on dynamic time warping. We propose an overview of the misalignment between the data corresponding to n different subjects. A detailed view focusing on the comparison between two of them can be obtain on demand. The proposed solution comes from a combination of signal processing and data visualization techniques. A web application implementing this proposal completes the contribution of this work.