AbstractSelf-supervised pretraining followed by supervised fine-tuning has seen success in image recognition, especially when labeled examples are scarce, but has received limited attention in
medical image analysis. This paper studies the effectiveness of
→