Multistage segmentation model and SVM-ensemble for precise lung nodule detection

Abstract

Lung cancer detection at its initial stages increases the survival chances of patients. Automatic detection of lung nodules facilitates radiologists during the diagnosis. However, there is a challenge of false positives in automated systems which may lead to wrong findings. Precise segmentation facilitates to accurately extract nodules from lung CT images in order to improve performance of the diagnostic method.