Neural networks with hybrid morphological/rank/linear nodes and their application to handwritten character recognition

We propose a general class of multilayer feed-forward neural networks where the combination of inputs in every node is formed by hybrid linear and nonlinear (of the morphological/rank type) operations. For its design we formulate a methodology using ideas from the back-propagation algorithm and robust techniques to circumvent the non-differentiability of rank functions. Experimental results in a problem of handwritten character recognition are described and illustrate some of the properties of this new type of nonlinear systems.