Comparison of type-2 fuzzy integration for optimized modular neural networks applied to human recognition

Abstract

In this paper optimization techniques for Modular Neural Networks (MNNs) and their combination with a granular approach is presented. A Firefly Algorithm (FA) and a Grey Wolf Optimizer (GWO) are developed to perform modular neural networks (MNN) optimization. These algorithms perform the optimization of some parameters of MNN such as; number of sub modules, percentage of information for the training phase and number of hidden layers (with their respective number of neurons) for each sub module and learning algorithm. The modular neural networks are applied to human recognition based on face, iris, ear and voice. The minimization of the error of recognition is the objective function. To combine the responses of the modular neural networks different type-2 fuzzy inference system are proposed and a comparison of results is performed.