Adam Red Panda Optimization for Detection and Severity Level Classification for Lung Cancer using CT Image

Abstract

Lung cancer is the leading cause of death worldwide, estimated to give rise to almost 7.6 million deaths annually. Early diagnosis is crucial in order to minimize fatalities associated with lung cancer. Two major imaging tests are Computed Tomography (CT) scans and chest X-rays, which are useful in diagnosing lung cancer. One of the major causes of death in the world today is lung cancer, and early detection and treatment is the key to successful management. Traditional methods of diagnosis, such as CT scans, have problems of accuracy and efficiency. This paper introduces a novel deep learning model, ARPO NasNet, which applies Adam Red Panda Optimization (ARPO), a novel optimization technique developed by the authors, to achieve better performance in detecting and classifying the severity of lung cancer based on CT scans. RPO (Red Panda Optimization) is a recently developed metaheuristic algorithm inspired by the behavioral characteristics of red pandas, which enhances the optimization process. The proposed method involves preprocessing CT images with median filters to remove noise, Deep Fuzzy Clustering (DFC) for segmentation of lung lobes, and Local Gradient Patterns (LGP) for feature extraction. The ARPO algorithm optimizes the NasNet model, improving its classification accuracy, precision, recall, and F1-score, thereby outperforming state-of-the-art methods. The proposed methodology demonstrates significant breakthroughs in lung cancer detection and the grouping of its severity phases, offering a solution for early and accurate diagnosis of lung cancer. Such results suggest the potential of ARPO_NasNet in clinical applications for the detection and treatment of lung cancer.