Fluid-Based Image Warping
In this thesis we present a technique of image deformation using dynamic
of fluids. We call this technique fluid warping. This technique was present
in first time by Jos Stam in
Stable Fluids, as one aplication of fluid simulation.
The main advantage of this approach is the effect is obtained in
the deformation of image, like a liquid image.
The key idea is to think of the image domain as a two-dimensional
incompressible fluid, and to use the Navier-
Stokes equations to model the fluid.
We deform the image through a vector field generated by equations.
The challenge now is to control the deformation.
For this purpose, we present three techniques for controlling the
fluid warping:
The first technique uses a constant viscosity and forces as parameters
control.
The second technique againg uses viscosity as a control tool, but in this
case it is variable in space and defined through of a image.
Finally, the third technique, more complex than the previous ones, need
three processes.
- Mark deformation goal by keyframes,
- employ the adjoint method to calculate derivatives
- and use the derivatives in a optimization process to find the necessary forces to achieve the goal of deformation.