Detection of Cancer Cells in Tissue Histopathology Images Using a Convolutional Neural Network

Cancer is a chronic disease with a high increase in the number of cases worldwide. In women, breast cancer is the most common cancer. One way to detect breast cancer is through histopathologic images of the tissue. Currently, histopathology detection is done manually, so it has shortcomings in the form of human limitations in classifying large numbers of images. In this research, a Convolutional Neural Network (CNN) model is developed to detect cancer cells in histopathology images automatically. The dataset used to train the model is 120,000 histopathology images labeled 1 (cancer) and 0 (normal) in the same ratio. Training the model is done with a batch size of 128 and epochs of 30. The confusion matrix and AUC-ROC method measure accuracy. From this study, the CNN model proved accurate in predicting tissues with cancer cells with a precision value of 94.76 percent, recall of 98.07 percent, and F-1 score of 96.39 percent.