Self-supervision for medical image classification: state-of-the-art performance with ~100 labeled training samples per class

1 code implementation • 11 Apr 2023 • Maximilian Nielsen, Laura Wenderoth, Thilo Sentker, René Werner

Exploiting the learned image representation, we achieve state-of-the-art classification performance for all three imaging modalities and data sets with only a fraction of between 1% and 10% of the available labeled data and about 100 labeled samples per class.

Image Classification Medical Image Classification