Automated aperture filter design by stochastic optimization

Aperture filters are a relatively new class of image operators, where in addition to the domain constraint, imposed by windowing, the range constraint is defined, effectively condensing probability mass to improve filter output estimates. In this paper the automatic design methodology by means of stochastic optimisation is proposed, which can generate an aperture filter to suit a specific application. Usability of presented method was verified experimentally, and the results are presented in the paper.