Review of Semi-Supervised Learning Methods for Medical Computer-Aided Diagnosis Systems
Bohdan V. Chapaliuk, Yuriy P. Zaychenko · 2020
Building computer-aided diagnosis systems remain one of the biggest challenges because of the complexity of the medical data and lack of massive labeled dataset. Semi-supervised methods provide the framework to reuse available unlabeled data to improve model accuracy and generality. This article considers two main strategies that are used in medical application to work with unlabeled data: label propagation and representation learning. Our review shows how these approaches used in the medical application and which modern strategies can be used to apply them.