Skip to Main Content
A new method is proposed to recognition motion type from long motion data. Since a high dimensionality represents original motion data, so it is difficult to find difference among motion data. Here we project motion data onto low dimensional space, and then we can separate different motion type based on low dimensionality. Experiments test our method and compare the performance with that of other methods.
Control Conference (CCC), 2010 29th Chinese
Date of Conference: 29-31 July 2010