Fast efficient computation of the discrete cosine transform (dct) of digital video data

摘要

A method for selectively computing and quantizing the components of the Discrete Cosine Transform of image data is presented. The method employs decomposition of the image data to successively obtain functions possessing even or odd symmetry about axes of the data. Threshold conditions may be applied to these auxiliary functions to determine which transform components may be deemed to be zero without further computation. Considerable reduction in the number of computations and memory required to compress video data is achived.