Fast software image stabilization with color registration

We present the formulation and implementation of an image stabilization system capable of stabilizing video with very large displacements between frames. A coarse-to-fine technique as applied in resolution and in model spaces. The registration algorithm uses phase correlation to obtain an initial estimate for translation between images; then Levenberg-Marquardt method for non-linear optimization is applied to refine the solution. Registration is performed in color space, using a subset of the pixels selected by a gradient-based subsampling criteria. This software implementation runs at 5Hz on non-dedicated hardware (Silicon Graphics R10000 workstation).