SpiLenet based detection and severity level classification of lung cancer using CT images

Abstract

Lung cancer is the type of cancer, which causes the global mortality. However, predicting and testing remains a serious issue due to its widespread and rapid growth. Hence, this research proposed the SpiLenet for lung cancer detection using computed tomography (CT) scan images. Initially, CT images are taken from a specific dataset, which are pre-processed by Savitzky-Golay (SG) filter. Then, the lung lobe segmentation is performed by Dense-Res-Inception Net (DRINet). Following that, the identification of lung nodule is carried out through a grid-based approach. Feature extraction (FE) is performed to extract key features for further analysis. Finally, lung cancer detection is conducted using SpiLenet, a model created by combining SpinalNet and LeNet. Experimental results demonstrate that SpiLenet achieves an accuracy of 92.10 %, an F-measure of 90.40 %, and a precision of 91.10 %.