System and method for transforming and filtering a video image

摘要

A video transformation and filtering system generates for each source image data point a plurality of target image addresses indicating a displacement between a point at which the source image data point maps into the target image and each data point of the target image which is affected by the source image data point. By providing either sequentially or in parallel a separate processing for each target image data point that is affected by a current source image data point values for all of the target image data points can be generated from a single pass of the source image data points. By effectively providing the filtering in the domain of the target image instead of the source image a single filter function with a fixed number of points in the target image domain can be used to attain low pass filtering of the resulting target image irrespective of the degree of size compression of the video image.