Computer System Prototype for Qualitative and Quantitative Evaluation of Selected Movement Activities
Tomasz Hachaj, Marek R. Ogiela · 2019
The aim of this paper is to propose prototype of a computer system for complex analysis of motion kinematic including data acquisition, processing, classification to one of the previously created classes and finally analysis based on comparison of input signal to motion template. We have used wireless inertial measurement unit sensors based costume, Gesture Description Language motion segmentation module, Hidden Markov Models for classification, and Dynamic Time Warping for comparison of acquired motions to actions templates. We have evaluated the proposed solution on karate motions dataset containing 8 motion classes. We have obtained total recognition rate 0.92. The analysis done by an algorithm has been evaluated by an expert. All differences indicated by our method between templates and recordings were approved by an expert and judged as meaningful.