An Efficient Approach to Classify Lung Cancer Tissue Cells Using Transfer Learning Techniques
J. Jyostna, M. Santhoshi, S. Sailaja · 2023
One of the deadliest malignancies in the world is lung cancer. In this paper, we are using histopathological images of lung tissue to classify them into cancerous and non-cancerous. We proposed different transfer learning approaches like VGG16, ResNet50, and DenseNet121 to classify the images of tissues into healthy and two types of cancerous tissues. Of the three models, the ResNet50 and DenseNet121 provided better accuracies in training and testing almost reaches to 96%. VGG16 provided 91% accuracy because its complex layers take a long time to be trained. With a high computational capacity system and better training, this model can perform much better. All three proposed models gave good accuracy compared to the basic CNN model and KNN model.