Collaborative Federated Learning in Healthcare Systems

Abstract

Collaborative Federated learning in the health sector can be viewed as a distributed privacy-aware model in which data is not centralized and it enables N clients to build a Deep Learning model or a Machine Learning model using decentralized data. One major benefit of collaborative federated learning is that it allows organizations to train models on large data sets than they could on their own. How machine learning is used in the healthcare industry can be increased better by improving patient care and outcomes by using collaborative federated learning. In the authors for the segmentation of brain tumours propose a federated learning architecture called SU-Net and the 'Brain MRI Segmentation' dataset from Kaggle is the dataset used for the study. Medical image steganography deals with hiding the secret medical image within the public or private cover image which is significant in the healthcare sector as everyone was relying on teleradiology applications during the covid scenario.