In the field of signal processing we frequently encounter the
use of transforms. Transforms are named such because they
take a signal and transform it into
another signal, hopefully one which is easier to process or
analyze than the original. Essentially, transforms are used to
manipulate signals such that their most important
characteristics are made plainly evident. To isolate a
signal's important characteristics, however, one must employ a
transform that is well matched to that signal. For example,
the Fourier transform, while well matched to certain classes
of signal, does not efficiently extract information about
signals in other classes. This latter fact motivates our
development of the wavelet transform.