
Chapter 15
Wavelets
15.1 Waves and Wavelets
We have seen in Chapters 7 and 14 that the discrete Fourier transform and its cousin, the
discrete cosine transform, are of enormous use in image processing and image compression.
But as we saw with the Fourier transform, this power does not come without cost. In
particular, as the Fourier transform is based on the idea of an image being decomposed into
periodic sine or cosine functi ons, we have to assume some sort of periodicity in the image
to make the theory fit the practice. This periodicity was obtained by assuming that the
image “bent round” to meet itself at the top and bottom, and at the left and