Radon transform-based improved single seeded region growing segmentation for lung cancer detection using AMPWSVM classification approach

Computer-Aided Diagnosis is a safe diagnostic procedure that uses CT scan images for the early detection of lung cancer. The quality of the CT scan images can be enhanced using digital image processing tools to assist physicians. A lot of researchers had worked on enhancing the quality and classification of lung cancer CT images. The problem of choosing the appropriate prediction algorithm for the classification of CT images and the proper decision support system remains a major task in the research field. Here, in this research article, the experimental analysis work CT images are taken from both Public & In-house clinical lung cancer Images. The proposed work was carried out in four phases. The first stage is to remove noise from the input image and enhance the contrast of the image's anomalous region by using Hybrid CLAHE with Unsharp Masking Technique. The second stage involved segmenting the denoise image and applying the Radon Transform Based Improved Single Seeded Region Growing segmentation method to accurately detect the lung tumor. A third stage entails extracting the GLRLM & GLCM features from the segmented ROI image. Finally, images of benign and malignant lung cancer are classified using the Advanced Marine Predator algorithm with SVM (AMPWSVM) Classifier. Using the AMPWSVM classifier an accuracy of 93.3% is obtained for the Public LIDC database and an accuracy of 90% is obtained for In house clinical dataset.