MOTION COMPENSATION USING 2ND-ORDER GEOMETRIC TRANSFORMATIONS

Motion Compensation is a very effective process which can be used in V ideo Data Compression for increasing the prediction accuracy in inter- frame coding. The conventional way of performing motion compensation a s defined in the H.261 and MPEG standards, is to divide the ''new'' fr ame into a number of nonoverlapping rectangular blocks and then find t he best match for each of these blocks by displacing a search window o f the same size in the ''previous'' frame (or also the next frame for bidirectional prediction in MPEG). A block prediction based on this me thod is only accurate if the motion between the two frames is of a tra nslational nature. There is no flexibility to change the size, rotatio n, or shape of objects if the motion that has taken place between the two frames contains such distortion, In this paper we evaluate the use of second-order Geometric Transformations (G.T.'s) as a means of:moti on compensation. These transformations are capable of modeling the kin d of changes mentioned above and therefore when used properly they hav e the potential of providing a much more accurate prediction during mo tion compensation.