Segmentation of Lung Nodule in CT Data Using K-Mean Clustering
Lung cancer is a rapidly growing deadly disease among human beings. Lung cancer occurs because of smoking,
radon gas, air pollution, asbestos,and genetics. The disease can be cured if detected in early stage. The proposed algorithm in
this work performsfiltering of CT images followed by morphological operations to extract actual lung region in an image.K-
means segmentation is performed to detect probable cancerous areas. Features of this probable area are extracted to perform
classification using support vector machines as normal, moderate or severe. The results of the method are encouraging to
have an automated method for detection.
Index Terms- Computed Tomography, K- mean clustering, SVM