%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2016/07.18.18.27
%2 sid.inpe.br/sibgrapi/2016/07.18.18.27.01
%@doi 10.1109/SIBGRAPI.2016.061
%T Using 3D Texture and Margin Sharpness Features on Classification of Small Pulmonary Nodules
%D 2016
%A Felix, Ailton de Lima Filho,
%A Ferreira Junior, José Raniery,
%A Oliveira, Marcelo Costa,
%A Machado, Aydano Pamponet,
%@affiliation Universidade Federal de Alagoas
%@affiliation Universidade de São Paulo
%@affiliation Universidade Federal de Alagoas
%@affiliation Universidade Federal de Alagoas
%E Aliaga, Daniel G.,
%E Davis, Larry S.,
%E Farias, Ricardo C.,
%E Fernandes, Leandro A. F.,
%E Gibson, Stuart J.,
%E Giraldi, Gilson A.,
%E Gois, João Paulo,
%E Maciel, Anderson,
%E Menotti, David,
%E Miranda, Paulo A. V.,
%E Musse, Soraia,
%E Namikawa, Laercio,
%E Pamplona, Mauricio,
%E Papa, João Paulo,
%E Santos, Jefersson dos,
%E Schwartz, William Robson,
%E Thomaz, Carlos E.,
%B Conference on Graphics, Patterns and Images, 29 (SIBGRAPI)
%C São José dos Campos, SP, Brazil
%8 4-7 Oct. 2016
%I IEEE Computer Society´s Conference Publishing Services
%J Los Alamitos
%S Proceedings
%K lung cancer, small nodules, early diagnosis, computer-aided diagnosis, texture features, margin sharpness features, classification, machine learning.
%X The lung cancer is the reason of a lot of deaths on population around the world. An early diagnosis brings a most curable and simpler treatment options. Due to complexity diagnosis of small pulmonary nodules, Computer-Aided Diagnosis (CAD) tools provides an assistance to radiologist aiming the improvement in the diagnosis. Extracting relevant image features is of great importance for these tools. In this work we extracted 3D Texture Features (TF) and 3D Margin Sharpness Features (MSF) from the Lung Image Database Consortium (LIDC) in order to create a classification model to classify small pulmonary nodules with diameters between 3-10mm. We used three machine learning algorithm: k-Nearest Neighbor (k-NN), Multilayer Perceptron (MLP) and Random Forest (RF). These algorithms were trained by different set of features from the TF and MSF. The classification model with MLP algorithm using the selected features from the integration of TF and MSF achieved the best AUC of 0.820.
%@language en
%3 PID4357869.pdf