Non-linear parametric derivation of contour detectors for cellular images

Abstract

This paper proposes an extension of the widely used Logarithmic Image Processing (LIP) models by means of parametrization. The mathematical structure of the mentioned models is that of cone or vector space. Once the mathematical investigation defined the boundaries of these structures, parametric extensions of the known models are straight-forward. It has been showed that the implementation of Laplacian edge detector techniques under the LIP model yields superior performance. In this paper we shall prove that parametrization not only adds flexibility, but may also lead to superior quality.