A Federated Learning-Based Approach for Classification of Histopathology Images

Nowadays, deep learning and machine learning methods are widely used in computer-aided diagnosis. In healthcare studies, they provide reliable and effective solutions for diagnosis and detection. However, important problems such as ensuring the privacy of patient data arise in these studies. In this study, the HAM10000 dataset was used, and the classification of histopathological images was conducted using the Federated Learning (FL) approach and deep learning methods to ensure data privacy and security. As a result of the tests performed, the classification accuracy rate of the proposed method was observed to be 82.04%.