Figure 6From: Learning motion primitives and annotative texts from crowd-sourcing A human full-body motion is converted to multiple descriptive sentences. This conversion make it possible for a humanoid robot to understand observation of the human motion in the forms of sentence.Back to article page