GENERALIZED IMAGE WARPING USING ENHANCED LOOKUP TABLES

Image warping includes a wide variety of algorithms used in the spatia l transformation of images. Warping typically involves a geometric tra nsformation governed by a set of polynomial equations followed by some form of interpolation. When the geometric transformation does not cha nge from one image to the next, or a more generalized warp is desired, a spatial lookup table (LUT) composed of precalculated inverse (or fo rward) mapping coordinates has been used instead of performing the sam e geometric transformation on every image. This article presents an al gorithm that takes the LUT approach one step further. It uses an enhan ced LUT (ELUT) that incorporates precomputed data transfer and interpo lation information. Once computed, it speeds up processing by making d ata transfers more efficient and eliminating the need for the pixel ad dress and interpolation coefficient calculations. This article also de scribes the implementation of ELUT-based image warping on the high-per formance TMS320C80 Multimedia Video Processor (MVP). To produce an int eresting and difficult generalized warp on a 512 x 512 8-bit gray-scal e and 32-bit color image, it takes only 15.1 and 32.4 ms, respectively . (C) 1998 John Wiley & Sons, Inc.