We present a novel technique for motion annotation that adapts to a person's style and vocabulary of basic movements (gestures). The system segments continuous motion sequences into gestures, which it then documents in a personalized annotation with an intuitive hierarchical representation. Initial testing suggests that software based on this technique could be an effective teaching aid for dance and sports.
ASJC Scopus subject areas
- Signal Processing
- Media Technology
- Hardware and Architecture
- Computer Science Applications