A comprehensive review on federated learning based models for healthcare applications.

Publication date: Dec 01, 2023

A disease is an abnormal condition that negatively impacts the functioning of the human body. Pathology determines the causes behind the disease and identifies its development mechanism and functional consequences. Each disease has different identification methods, including X-ray scans for pneumonia, covid-19, and lung cancer, whereas biopsy and CT-scan can identify the presence of skin cancer and Alzheimer’s disease, respectively. Early disease detection leads to effective treatment and avoids abiding complications. Deep learning has provided a vast number of applications in medical sectors resulting in accurate and reliable early disease predictions. These models are utilized in the healthcare industry to provide supplementary assistance to doctors in identifying the presence of diseases. Majorly, these models are trained through secondary data sources since healthcare institutions refrain from sharing patients’ private data to ensure confidentiality, which limits the effectiveness of deep learning models due to the requirement of extensive datasets for training to achieve optimal results. Federated learning deals with the data in such a way that it doesn’t exploit the privacy of a patient’s data. In this work, a wide variety of disease detection models trained through federated learning have been rigorously reviewed. This meta-analysis provides an in-depth review of the federated learning architectures, federated learning types, hyperparameters, dataset utilization details, aggregation techniques, performance measures, and augmentation methods applied in the existing models during the development phase. The review also highlights various open challenges associated with the disease detection models trained through federated learning for future research.

Concepts Keywords
Biopsy Cancer detection
Cancer Covid-19 prediction
Extensive Deep learning
Models Disease detection
Pneumonia Federated learning
Machine learning
Multi-disease classification
Pneumonia detection

Semantics

Type Source Name
disease MESH causes
disease MESH pneumonia
disease MESH covid-19
disease MESH lung cancer
disease MESH skin cancer
disease MESH Alzheimer’s disease
disease VO effective
disease MESH complications
disease VO effectiveness
disease MESH Cancer

Original Article

(Visited 1 times, 1 visits today)