RESUMO
We discuss and implement computer approximations of fractal and multifractal hypersurfaces. These hypersurfaces consist of reconstructions of a stochastic process in the real space from randomly distributed variables in the discrete wavelet domain. The synthetic surfaces have the usual fractional Brownian motion as a particular case, and inherit the correlation structure of these fractals. We first introduce the one-dimensional version of these surfaces that obey a weak self-affine symmetry. This symmetry appears in the wavelet domain as a condition on the second moments of the probability distributions of the wavelet coefficients. Then we use these relations to define the fractals and multifractals in d dimensions. Finally, we concentrate on the generation of samples of these hypersurfaces.