Affine Image Warping on Programmable Graph Architecture using Grid Space Algebra

We introduce a new model for graphics texture mapping and image transformation called the grid space algebra model. This model has been designed to facilitate the development of a wide range of graphics applications on a novel multimedia processor architecture that is based on transforming matrix groups to other groups such as permutation group or to a programmable graph architecture which is based on the isomorphism between groups and Cayley graphs. Such architectures have a great potential for speeding up graphics operations by not only exploiting their inherent parallelism as done in stream based architectures but also expediting the operations by mapping them to a domain that does not require expensive matrix operations. We give an algorithm for affine transformation of images and affine texture mapping using the grid space approach. Specifically, we emulate the FLC algorithm for texture mapping using this new approach.