AN INVESTIGATION OF THE NATURE OF PARAMETERIZATION FOR THE HOUGH TRANSFORM

A novel parameterization method for the Hough transform is reported. I nstead of the conventional non-parametric form, the parametric form is used and copies of the transformed shape are plotted on two-dimension al slices of the Hough space. It is shown that the corresponding param eterization has uniform precision with respect to translation, and can cels out the quantization uncertainty due to image digitization. A pro blem of the Hough transform is discovered which is due to non-uniform discretized voting. It is shown that the above class of parameterizati ons avoids the problem. Finally, a particular solution of the paramete rization scheme is described which is called the Fourier parameterizat ion. It is shown that the parameterization has uniform precision with respect to the affine transformation. (C) 1997 Pattern Recognition Soc iety.