Abstract— Lung cancer is the most dangerous disease. The chest X-ray (CXR) is the most widely used and crucial imaging method for detecting lung cancer. The CXR is chosen due to its accessibility, affordability, non-invasiveness, and ease of acquisition. Machine Learning (ML) is used to automate lung cancer identification. ML in medical imaging has the potential to reduce medical professionals’ burden throughout the diagnostics and screening process. Lung segmentation is an important step in lung analysis. Several factors that make lung segmentation difficult are: 1) The size and structure of the lungs change as a result of gender, age, and heart size. 2) Opacity caused by severe pulmonary disease with a high-intensity value. 3) The patient’s garments or medical gadgets obscure the entire visibility of the enclosed foreign entity, like the lung field. This study tries to address all of these issues. Graph cut and Fuzzy C-means (FCM) are two segmentation algorithms employed. The segmentation result from both techniques is fed into an ML model for classification followed by feature extraction. The ML model attempts to identify lung images as “normal” or “cancer.” The outcome of the ML model followed by both segmented techniques is compared. The comparison results show that the FCM delivers more exact results on screening lung cancer.

Keywords— Lung, Segmentation, Fuzzy, Graph Cut, Support Vector Machine.