ABSTRACT
Geometric transformations allow the removal of geometric distortion that happens when an image is captured. There are two types of geometric transformation: pixel coordinate transformation and brightness interpolation. Brightness interpolation involves nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation. Bicubic interpolation does not suffer from the step-like boundary problem of nearest neighborhood interpolation and copes with linear interpolation blurring as well. The correspondence is established via the spatial transformation mapping function to assign the output point onto the input image. All possible simple mapping or transformations are special cases of affine mapping. Since the general affine transformation is characterized by six constants, it is conceivable to express this transformation by determining the new output image locations of any three input image coordinate pairs. In general, several points are estimated and a least squares technique is used to find the finest fitting transform. Nonlinear mapping involves twirl, ripple, and spherical transformation.
