MOTION SYNTHESIS FROM THE SEMANTIC SIGNALS

Abstract

It is still an open problem to reuse the motion capture data in an intuitive way. In this paper, we present a novel technology to synthesize animations from the low-dimensional semantic signals. The semantic signals are defined as the meanings which are visible to the animators, such as the angles of joints rotation around axis, the trajectories of joints, or other intuitive motion signals. A linear time-invariant system is used to model the relationship between the input semantic signals and the output human motions. Once the model parameters are estimated, the outputs of system can be effectively controlled by the inputs. Because the semantic signals can be intuitively sketched or modified by animators according to the general knowledge in everyday life, this may provide an intuitive tool for animators to create new animations from the existing data. Furthermore, a novel algorithm is also proposed to edit the semantic signals for creating the full body motions while considering the correlations among joints. This is different from the traditional methods which consider human joints as independent from each other.