Machine learning and deep learning have demonstrated significant promise for many kinds of medical imaging applications, including segmentation, classification, and detection. The quantity of data needed for developing effective models for medical images, however, is substantial and can be restricted by privacy restrictions like HIPAA and GDPR. Federated learning has gained popularity in the field of medical imaging as a privacy-focused approach. This meta-analysis, which focuses on current work, extensively investigates the use of federated learning in medical image processing. On images of the liver, stomach, colon, prostate, breast, and lungs, we examine methods employed throughout the previous five years. The paper explores techniques for classifying medical images, addresses data diversity and privacy concerns, and examines how federated learning is impacted by non-uniform(non-IID) data distribution. The Benefits and challenges of using federated learning for medical imaging are also discussed.