3D point cloud deformation based on moving least squares and control curves

A deformation method for 3D point cloud model is proposed based on moving least squares and control curves. The user controls the deformation by manipulating a set of curves. We set the critical points according to the topology information, and then make the curve fitting over the points using cubic spline interpolation. We can handle the contour of deformed object accurately by the contour fitting. Locally, the deformation takes the form of a rigid transformation, more accurate deformation than points-set controls. And it is applicable to deformation in arbitrary dimensions. Experiments show that 3D point cloud deformation and contour information fitting can be realized by this method and the deformation result is realistic.