AI-Enhanced Lung Cancer Detection Using the ResNext50 Architecture

Lung cancer as the first cause of cancer death among people in many countries shows the necessity of creating a straightforward as well as accurate test for identifying the disease at an early stage. In the last several years automatic learning (AI) of machine learning (ML) technologies have been regarded as great techniques for medical diagnostics with the help of imaging analysis. This article will take the ResNext50 architecture, a variant of convolutional neural networks (CNN), as a model. It will use the ResNext50 architecture as an example of the AI-enhanced lung cancer detection neural network. The methodology has the following steps: obtain a versatile database of CT scans (chest) with labeled cancerous lung lesions. As preprocessing steps, we apply techniques like normalization, resizing, and augmentation for standardization, and the augmentation for enhancement of the dataset. ResNext50 is embraced as the main backbone for the lung cancer detection system being an advantage of transfer learning and fine-tuning, which helps improve the model’s efficiency. We assess the performance of the model according to a set of metrics which estimate the correctness of the model. These include accuracy, sensitivity, specificity, precision, F1-score, and the area under the receiver operating characteristic (AUC-ROC) curve. Competitive evaluations are executed to compare the baseline methods and the superiority of the considered procedure. The results confirm that the ResNext 50 model reached a high level of 92%, with an accurate evaluation of metrics that matter. Performance check of the ResNext50 model versus baseline methods proves this model’s superiority and bringing out more significant advantages while working with advanced CNNs is recommended for lung cancer detection.