The Best Basis and Local Discriminant Bases Algorithm on Moderate Parallel Mimd Architectures

Introduction PACT Abstract Wavelet packets represent a generalization of the method of multiresolution decomposition and comprise the entire family of subband coded (tree) decompositions. Wavelet packet best basis selection has become a popular method in image compression, whereas the local discriminant bases algorithm is a relatively new feature extraction method based on a wavelet packet expansion. Both methods are very similar from an algorithmic point of view-this paper introduces a subband based parallelization which overcomes most of the diiculties of a straightforward parallel version of the sequential algorithms. Beside the higher eeciency the algorithm is easier to implement than its classical version.