Figure 2From: Learning motion primitives and annotative texts from crowd-sourcing A video containing human activity is openly available. Any user can assign the descriptive annotation to the video on the Internet.Back to article page