A Multiclass Classifier for Lung Cancer Diagnosis Using Resnet50

Lung cancer stands as a prominent cause of human mortality globally, with an annual fatality rate of approximately five million cases. Remarkably, this surpasses the combined death tolls of prostate and bone cancer. Despite its significant impact, early detection and diagnosis play a pivotal role in enhancing survival rates. Various diagnostic modalities are employed to assess lung cancer, with computed tomography (CT) scans providing crucial information about lung infections. The primary objective of this model is the precise identification and categorization of different lung cancer forms, namely large cell carcinoma(LCC), adenocarcinoma(ADC) and squamous cell carcinoma(SCC). The classifier is trained using a comprehensive dataset of CT scans, providing detailed input for the model. Leveraging the ResNet50 architecture and a robust multiclass classifier, this study achieves an impressive 97.14% accuracy in diagnosing lung cancer during testing.