Performance aspects of a novel neuron activation function in multi-layer feed-forward networks

Conventional single layer networks are limited by their inability to solve nonlinear classification problems. A modified neuron activation function has, recently, been proposed to extend the classification capabilities of single layer networks to cover some nonlinear problems. This paper shows that the classification capabilities of a multilayer network can also be improved by incorporation of the modified activation function, in that the required number of hidden layers and that of hidden neurons for a complex application can be reduced. A multilayer network, designed to perform textured image segmentation in computer vision applications is used in preliminary experiments to demonstrate the effectiveness of the new activation function.