Classificação de imagens tomográficas de ciência dos solos utilizando redes neurais e combinação de classificadores

Abstract

Pattern Recognition is a subject being used in a multidisciplinary scope, with different approaches. One of them is its application in computerized tomography images, commonly acquired in order to do medical diagnosis, but they have been used in several other applications as well, including Soil Science. The objective of this work is to study and to discuss the performance of neural network-based classifiers (Multilayer Perceptron and Radial Basis Functions) and classifier combiners (Bagging, Decision Templates and Dempster-Shafer) applied to identify materials in Soil Science multispectral images, acquired using Computerized Tomography. The results were evaluated by error estimation by Hold- Out and the Kappa coefficient.