Surface fitting and numerical gradient computations by discrete mollification (Q1963022)

The authors consider the problem of fitting of a surface, which is given only by noisy data (in \(\mathbb{R}^2)\). This is done by a so-called mollification procedure, which is a process of filtering and smoothing the data by a modified convolution process. Many results on the convergence of the method are presented, and also a large number of computed examples is illustrated by a total of 30 figures.