ABSTRACT
Wavelet packets are a natural and versatile extension of wavelet bases to extract useful and interesting information from signals. With the use of wavelet packets a very large class of bases can be constructed. In this type of transformation, the basis selection for a given application is made based upon a collection of functions. The wavelet-packet transforms offer more flexibility than the classical fast wavelet transform algorithm, albeit at the cost of a more computationally expensive algorithm. Wavelet packet generation uses the concept of graph-theoretic trees. In order to describe an algorithm to compute wavelet-packet transform, familiarity with the language of trees is necessary. There are two types of wavelet packets. These are the basic and general wavelet packets. Basic wavelet packets are a collection of wavelet and scaling functions, which are generated recursively. A more general form of wavelet packets are generated in turn by dilation and translation of the basic wavelet packets.
