Feature Flow Animation

Abstract

Existing feature flow animation techniques do not address texture aliasing or distortion when a critical point moves along some designated path. In this paper, we provide a smooth feature flow animation for planar domains. Feature flow behaviors are clearly shown through smooth animation achieved by configuring critical points and constructing spatio-temporal coherent patterns. Critical points can be seeded at any position in the flow domain. We partition the flow space into regions with Voronoi tessellation and construct the flow pattern in each region. When one critical point moves, we add the original vector field to the dynamic vector field, optimize the vector values at the boundary, and update flow patterns over time. Experiments demonstrate that our method can effectively address texture aliasing or distortion in feature flow animation.