Abstract
This research explores the use of explainable artificial intelligence techniques in diagnosis of prostate cancer. This study uses whole slide images for prostate cancer detection. The study uses various deep learning models such as CNN, ResNet50, VGG19, DenseNet, MobileNet, Xception for image classification. VGG19 outperformed all the other models based on the evaluation metrics. GRAD-CAM, an explainable AI technique was applied to VGG19 model to understand the outcome of the classification. The study's results are encouraging because they showed not only better accuracy but an interpretation of the result using XAI techniques. Following these disclosures, a methodical classification procedure was implemented, dividing the Gleason grades into their respective classes.
