Quantitative analysis of robot gesticulation behavior

Social robot capabilities, such as talking gestures, are best produced using data driven approaches to avoid being repetitive and show trustworthiness. However, there is a lack of robust quantitative methods that allow compare beyond visual evaluation. In this paper analysis performed compares two Generative Adversarial Networks based gesture generation approaches. The aim measure characteristics fidelity the original training data, but at same time keep track degree originality gestures. Principal Coordinate Analysis procrustes statistics new Fr\'echet Gesture Distance proposed by adapting Inception These three techniques taken together asses fidelity/originality generated