Harnessing Machine Learning for interpersonal physical alignment
Roi Yozevitch, Hila Zahava Gvirts, Ornit Apelboim, Elhanan Mishraky · 2018
This work presents a novel way to determine interpersonal physical synchrony state by inspecting hands' postures obtained from a unique 3D depth camera device named Leap-Motion Controller. Several ML methods are utilized such as SVM, shallow feed-forward ANN and XGBoot. We show that even a simple ANN can outperform XgBoost in simple classification tasks.