Valuation of climbing activities using Multi-Scale Jensen-Shannon Neighbour Embedding
Résumé
This paper presents a study carried out in a controlled environment that aims at understanding behavioural patterns in climbing activities. Multi-Scale Jensen-Shannon Neighbour Embedding [8], a recent advance in non linear dimension reduction, has been applied to recordings of movement sensors in order to help the visualization of coordination modes. Initial clustering results show a correlation with jerk, an indicator of fluency in climbing activities, but provides more details on behavioural patterns.
Origine : Fichiers produits par l'(les) auteur(s)