Seonghun Kim  Yangkyu Lim  Youngho Kim  Youngho Chai

10.5626/JOK.2019.46.10.989

Keyframe animation Motion Capture human motion authoring gesture reconstruction motion recording

Research on human motion perception has progressed along with the development of technology. The demand for the gesture recognition is increasing daily. In motion capture, data on each part of the body are acquired by using a sensor or camera in order to realize a natural movement as an actual motion. However, it is inefficient to use the sensor every time motion data must be obtained or to make repeat measures because of slight differences in motion. In addition, various problems must be solved in order to transfer the stored data to another person or modify it use for another action.
In this paper, we studied the trends of motion recognition research and analyzed characteristics of the motion reproduction method using keyframe animation and the Labanotation motion recording method. In addition, we proposed an action authoring method which addresses the disadvantages of the existing methods. By visualizing and recording the pattern of partial motions in the units of the images, it is possible to reconstruct other motions, including the meaning of the motion.