2021/01/04 by Niloofar Akhavan Javan, Javan, Niloofar Akhavan, Ali Jebreili +7
Medicine · #Radiomics and Machine Learning in Medical Imaging
paper · pdf · doi:10.48550/arxiv.2101.00948
Given that lung cancer is one of the deadliest illnesses, early\nidentification and diagnosis are critical to preserving a patient's life.\nHowever, lung illness diagnosis is time-intensive and requires the expertise of\na pulmonary disease specialist, subject to a significant rate of inaccuracy.\nOur objective is to design a system capable of accurately detecting and\nclassifying lung lesions and segmenting them in CT-scan images. The suggested\ntechnique extracts features automatically from the CT-scan image and then\nclassifies them using Ensemble Gradient Boosting methods. Finally, if a lesion\nis detected in the CT-scan image, it is segmented using a hybrid approach based\non Fuzzy Clustering and Level Set. To train and test our models we gathered a\ndataset that included CT images of patients residing in Mashhad, Iran. Finally,\nthe results indicate 96% accuracy within this dataset. This approach may assist\nclinicians in diagnosing lung abnormalities and avoiding potential errors.\n