Metrics for Motion Editing

Motion editing is increasingly being accepted as a means for expanding the range of motions available for animating characters with far less expenditure of time and effort than the original motion creation method. Adding flexibility and breadth to motion capture data is particularly valuable, provided the new motions are sufficiently “good.” Computable metrics for determining whether a motion is good enough for a certain context would enable the automatic generation of edited motions to be done with much greater confidence, vastly increasing the range of practical applications. In addition, a sufficiently powerful suite of metrics could aid in examining the strengths and shortcomings of different methods of motion editing. This research opens the field of metric analysis of motion editing and presents some first, and encouraging, results.