%0 Conference Proceedings
%T Segmentation of lung and its lesions in computer tomographic images
%D 2017
%8 Oct. 17-20, 2017
%A Colella, Sílvia Regina Leme,
%A Rittner, Letícia,
%@affiliation University of Campinas - UNICAMP
%@affiliation University of Campinas - UNICAMP
%E Torchelsen, Rafael Piccin,
%E Nascimento, Erickson Rangel do,
%E Panozzo, Daniele,
%E Liu, Zicheng,
%E Farias, Mylène,
%E Viera, Thales,
%E Sacht, Leonardo,
%E Ferreira, Nivan,
%E Comba, João Luiz Dihl,
%E Hirata, Nina,
%E Schiavon Porto, Marcelo,
%E Vital, Creto,
%E Pagot, Christian Azambuja,
%E Petronetto, Fabiano,
%E Clua, Esteban,
%E Cardeal, Flávio,
%B Conference on Graphics, Patterns and Images, 30 (SIBGRAPI)
%C Niterói, RJ
%S Proceedings
%I Sociedade Brasileira de Computação
%J Porto Alegre
%K medical image segmentation, interstitial lung diseases, computer tomography.
%X The purpose of this work is to propose two new automatic segmentation methods in CT images: one for the lungs and one for their lesions. The lung segmentation method uses morphological filters and the max-tree, a data structure that represents an image through its connected components. Results show that the method presented a good performance when compared to the manual segmentation and it was able to not exclude lesions located in the borders in most of the images, which is challenging when the lesions are small and disconnected located in this region. This method obtained an average Dice of 98%. The lesion segmentation method uses the image with the segmented lungs to calculate the features to train a classifier that distinguishes between normal tissue and abnormal tissue (which contains lesions). This method also presented good results as it turned out not being very sensible to parameters' choice and it obtained an average Dice of 62% for the slices with severe pathologies.
%@language en
%3 wtd-sibgrapi-2017-SilviaColella-camera-ready.pdf