Aplicação de conceitos de redes neurais convolucionais na classificação de imagens de folhas

Abstract

This work proposes an algorithm to solve one of the problems faced by taxonomists, the classification of leaves, a process that is done manually and takes a lot of time to accomplish, where the success rate depends simply on the professional capacity of the taxonomist. With the advance of technology and the emergence of deep learning, the convolutional neural networks were the focus of large studies in the area of image classification. Nowadays several researchers dedicate their time in developing algorithms for classification of foliar images based on the intrinsic concepts of convolutional neural networks. The developed neural network in this work was trained and tested on an image dataset available by the PlantCLEF project in the year 2016 among others.