A Novel Method for Lung Nodule Segmentation and Lung Cancer Severity Categorization using Deep Learning Models
A. Anto Sagaya Priscilla, R. Balamanigandan
Abstract
The fluctuating nature of lung cancer, which is impacted by recurrent radiation exposure and Computed Tomography (CT) pictures, makes the early detection of the disease challenging. Even seasoned professionals find manually inspecting the provided photos for lung nodules to be tiresome. In this research, a novel hybrid model that combines U-net and Dense Convolutional Network (DenseNet) - 121 is presented to enhance the diagnostic ability and accuracy in classifying the lung cancer disease. In comparison to the current methods, the suggested model for classifying lung nodules and lung cancer is assessed using metrics including accuracy, precision, sensitivity, and F1-Score. It is evident from the comparison results that the suggested model outperforms the current models.