Research On Automatic Generation Of Labanotation Based On Human Motion Capture Data

Labanotation is a kind of recording system for human dance. It can be used for choreography, dance education and dance recording. Motion capture is a kind of tech-nology that can be used to track the motion of human body precisely and record the mo-tion with human motion data in high dimensional space. People can easily understand the motion described in Labanotation, but can hardly notate human body motion with it. On the contrary, human motion capture data can be acquired conveniently with the ad-vanced motion capture devices, but the high dimensional data make it impossible for people to read and use it. Thus, if we can combine the motion capture technique with Labanotation, which means generating Labanotation from human motion capture data automatically, we can simplify the notation process of Labanotation. The aim of the subject that automatic generation of Labanotation based on human motion capture data is to make it easier to notate human motion with Labanotation so as to apply it to the recording, inheritance and protection for folk dances and operas-the art of body movement-in China.Research in this article was centralized in the method to convert human motion data into Labanotation characters. As two kinds of recording system for human motion, it is conceivable that there must be lots of similarities between human motion data and Labanotation. So, it is necessary to explore the principles of Labanotation and motion capture technology to determine the feasibility of this converting process before carry-ing on this research. Considering the consistency between human motion data and Labanotation, we can say that generating Labanotation from human motion data could be workable.Then, the first step to conduct this converting study is to parse the file format for human motion data. A parsing method based on tetrad description for human joints has been used to determine semantics of the joints in human skeleton defined in motion data file. And then, the position of each part of human body in world coordinate could be calculated via the transforming from rotation data in Euler Angle to position data.In this article, an approach to segment human motion sequence into distinct ele-ment motions was proposed for extracting unit motion for human movements. It is a type of clustering approach based on Laban direction spaces. Two methods, which based on kinematic feature and intrinsic dimensionality separately, were also used to segment human motion data in accordance with the motion of human body parts. Thus, human motion can be analyzed on the basis of element motions extracted from motion data. There were two methods applied to this analysis. The one used to analyze non-supporting motion is based on theory of Laban spaces while the one for analyzing supporting motion is rule-based. This analysis is to substitute the element motion seg-mentations from motion data with Labanotation characters.After all, a framework for converting human motion data into Labanotation was designed in this article. Furthermore, we developed a platform for generating Labanota-tion automatically from human motion capture data. It can be observed that the gener-ated Labanotation can describe key postures of the original human motion in motion data correctly after comparing them. So, we finally achieve the goal of simplifying the notation process for Labanotation.