Analysis of Convolutional Neural Network Architectures for the Classification of Lung and Colon Cancer

Cancer is the most common disease, which leads to death worldwide. Colon, prostate, breast, bladder, and lung are the common types of cancer among men and women. This work helps researchers to compose an automated classification of lung and colon cancer using histopathology images. This study compares various Convolutional Neural Network (CNN) architectures for image classification using the feature extraction method for the LC25000 dataset. In this method, a pre-trained network extracts features from the dataset, and the multi-class classification approach classifies the test images. All models show excellent performance in classification. This study achieves accuracy between 97.32% to 99.79% for lung-colon cancer image classification.