Style-Based Motion Editing

Both motion capture and motion editing are important human motion controlling methods: the former obtains the data captured from the real and natural human joints, and the latter provides the method to edit data. This paper combines the two methods using IK algorithm, and proposes a style-based motion editing approach. During the preprocessing phase, in order to take advantage of the motion capture data, the method performs a PCA on style motion to extract its style, and calculates the key frames of the base motion; then, during the on-line motion editing phase, the method transfers the style extracted from style motion to certain key frames of base motion using IK algorithm, afterwards, interpolates the edited key frames to generate a new motion with the styles of the two motions. The experiment results demonstrate that the style-based motion editing approach has the advantage of better synthesizing performance.