LAUSR

Abstract Lung cancer (LC) is one of the major causes of death worldwide. Early diagnosis helps to improve the patient survival outcome. The surgeon makes use of Computed Tomography (CT) for detecting LC using the aid of a Computer-Aided Diagnosis (CAD) system to identify LC effectively, but it has issues related to processing time and diagnostic precision that continue to pose significant challenges. To address this, a Deep Residual Xception Network (DRX-Net) approach has been introduced for identifying the LC. Initially, the CT image is obtained and then denoising is performed using a Wiener filter. Subsequently, the segmentation of lung nodule is conducted using Pyramidal Attention-based Y Net (PAY-Net), which uses a hybrid loss function combining Binary Cross Entropy, Tanimoto Similarity, and Dice Loss. The segmented image undergoes data augmentation followed by feature extraction. For LC detection, the selected features are processed using DRX-Net, which merges the Xception with a Deep Residual Network (DRN). Furthermore, the results show that the proposed DRX-Net achieved an accuracy of 93.988%, a True Positive Rate (TPR) of 95.567%, and a True Negative Rate (TNR) of 91.432% when evaluated using a K Group of 8.