Linear Dilation-Erosion Perceptron Trained Using a Convex-Concave Procedure

Mathematical morphology (MM) is a theory of non-linear operators used for the processing and analysis images. Morphological neural networks (MNNs) are whose neurons compute morphological operators. Dilations erosions elementary MM. From an algebraic point view, dilation erosion that commute respectively with supremum infimum operations. In this paper, we present linear dilation-erosion perceptron (\(\ell \)-DEP), which given by applying transformations before computing erosion. The decision function \(\ell \)-DEP model defined adding Furthermore, training can be formulated as convex-concave optimization problem. We compare performance other machine learning techniques using several classification problems. computational experiments support potential application proposed binary tasks.