Lung Cancer Detection Using Guided Filtering and SVM Classifier

One of the major reasons for a large number of deaths around the world is cancer. Among various kinds of cancers, lung cancer is responsible for a large number of deaths. Therefore, early detection of it is very necessary to save patients life, and provide required treatment. The reliance of conventional diagnostic techniques on radiologist expertise can cause errors and delays in diagnosis, and both CT scans and X-rays are often hindered. This research article proposes an automated method based on guided filtering for image enhancement, and a SVM classifier for lung cancer detection. These applications employ guided filtering to reduce noise and improve the edge details of lung CT images in order to get improved feature extraction. The processed enhanced images train an SVM classifier that achieves good separation of cancerous and non-cancerous regions using extracted features. Results of experiments show the proposed method achieves high classification accuracy in which its sensitivity and specificity are superior to standard methods. This method can also be used to clearly automatically and reliably detect lung cancer so that healthcare professionals would be able to early diagnose and improve patient's outcome.