Convolutional Model with Classification through Izhikevich Neuron

This paper presents the fusion of two paradigms of neural networks: the Convolutional Neural Networks from deep learning and the third generation Izhikevich neuron.This fusion has the purpose to replace the multilayer perceptron layers, that usually represent a computational cost and a large training time, for a paradigm created to classify with a single neuron.The experimentation is carried out in the classification domain, predicting the directions of rotation in a simulator of self-driven vehicles.The experiments show similar results to the multilayer perceptron model in the evaluation metrics but the training time is reduced.