Analysis of CT Scan Images to Predict Lung Cancer Stages Using Image Processing Techniques

• Main Authors: Islam, Mahmudul, Mahamud, Al Hasib, Rab, Raqeebir
• Format: Conference Proceeding
• Language: English
• Subjects: Classification; Computed tomography; Feature extraction; Feature Selection; Filtering; Gabor filters; Histogram Analysis; Image Processing; Image segmentation; Lung; Lung cancer; Lung Cancer Stage; Lung Nodule; Machine Learning; Morphological Opening; Segmentation; Texture Analysis; Thresholding
• ISSN: 2644-3163
• DOI: 10.1109/IEMCON.2019.8936175

Lung cancer is one of the most dangerous and common cancer diseases in the world. Early detection of lung cancer can increase survival time of a patient. It is difficult for doctors to identify the cancer stages from Computed Tomography (CT) scan images. In this era of technology computer-aided system can help us to predict lung cancer stages more accurately. Inspired by the recent success of image processing and machine learning techniques in medical field we have developed models using Gray level co-occurrence matrix (GLCM) based texture image analysis and Statistical parametric approach for helping doctors to detect lung cancer stages. Our approach involves image acquisition, preprocessing, feature extraction and finally classification. For feature extraction purpose two approaches are used: Gray level co-occurrence matrix (GLCM) based texture image analysis and Statistical parametric approach. For detecting lung cancer stages four different classifiers are used and obtained the highest accuracy 78.95% with 0.77 precision and 0.83 recall using Support Vector Machine(SVM) in the Statistical parametric approach of feature selection.