Estimation of the Number of Filters in the Convolution Layers of a Convolutional Neural Network Using a Fuzzy Logic System.

In this paper, we propose to search for the best number of filters in the convolution layer of a convolutional neural network, we used a fuzzy logic system to find the most suitable parameters for the proposed case study. In addition to this we make use of the Fuzzy Gravitational Search Algorithm method to find the parameters of the fuzzy system memberships.