Enhanced Lung Cancer Classification and Prediction based on Hybrid Neural Network Approach
A. Gopinath, P. Gowthaman, Lisa Gopal, Md. Abul Ala Walid, M Manju Priya, K Keshav Kumar
Abstract
Cancer is the leading killer on a global scale. As the leading cause of cancer-related mortality, lung cancer is among the most prevalent forms of the disease. Uncontrolled cell growth in the lung tissues is the hallmark of lung cancer, a potentially fatal malignancy. This uncontrolled growth can spread to nearby tissues and other sections of the body if the disease is not treated in its early stages. Because there are few or no signs in the early stages of this disease, most cancer cases are typically detected in the later stages, making early detection particularly challenging. Survival rates can be increased via early detection and treatment of lung cancer. And early detection of lung cancer is a crucial step in this direction. Greyscale conversion, image scaling, and noise removal are all part of preprocessing once the input image has been provided. The gaussian filter is used to improve the quality of an image. Following segmentation using the watershed threshold method, features are extracted using area, perimeter, and eccentricity, and the model is trained using an enhanced GAN-mask region-based CNN. In this proposed approached a novel model the GAN-R-CNN model that combines the GAN with the CNN technique.