VGG BASED MULTICLASS CLASSIFICATION FOR LUNG CANCER

Abstract

Abstract— Lung cancer is the most common cause of cancer-related cancer, having been estimated to claim about 9 million lives worldwide primarily owing to the diseases symptomatology often presenting itself in its advanced stage. Early detection is essential in the management of patients to enhance their wellbeing. Detecting lungs cancer is not challenging but differentiating its stages is quite difficult task. Traditional diagnostic methods, although efficient, occasionally contradict themselves as to the issue of precise and cost effective, especially at the onset of the disease. Among these, the most used technique is Computed Tomography. However, the Based on these scans, its interpretation can be difficult and sometimes may be subjective in nature which in turns creates variability in the levels of diagnosis accuracy achieved. To address these challenges, the deep learning approaches, particularly in (CNNs), have recently been used in the subsequent automation and the improvement of the analysis process of medical images. Our subsequent approach, training on VGG-Net and CT scan data to classify it into normal, benign and malignant cases. By leveraging in this model, we hope to get better accuracy and results consistency in detection of lung cancer than the previous techniques of diagnosis. Our model achieved 99. 2% accuracy, with both precision, recall, and F1 score showed value of 1. 0.