Research On Human Motion Editing Based On Motion Capture Data

With the popularity of motion capture technology and the rapid development of motion capture equipment since 1990s, computer animation has made great progress, and has been widely used in many fields, such as game, film, medicine and military. Problems, however, occur concerning how to produce animation efficiently and effectively using motion capture data. Under this background, human animation based motion capture data has been a hot topic in Computer Graphics recently.The work of this thesis is to study the technology of the human motion editing based motion capture data, aimed for exploring the motion editing methods, which are more suitable for human motion features and more in line with the human visual requirements of continuous.After describing the human skeleton hierarchical skeleton model and human motion, the format of present motion capture data are analyzed. Pose interpolation is used to create smooth transition of the motion segments. The thesis makes use of quaternion to compute the joint's rotation in every frame in the human motion data and produce the spherical linear interpolation between the two pose. A method of motion blending on the basis of the time warping is proposed. The processing of motion blending is divided into motion synthesizing and action inlay. The traditional dynamic time warping algorithm is improved by uniform scaling the time series and a lower bounding technique to increase the feasibility and efficiency of algorithms. A motion path transforming method based on extracting the motion feature is proposed. Original motion path is divided up to several short periodic motion paths; each periodic path is space-transformed to a new periodic path, with adjusting the offset of the direction .After all short periodic motion paths are transformed, the motion along the new path is generated.