Semi-Supervised Relational Contrastive Learning
This work addresses the challenge of reducing reliance on costly expert labeling for medical image diagnosis, offering a semi-supervised approach that is incremental in combining existing techniques for improved data utilization.
The paper tackles the problem of disease diagnosis from medical images by introducing Semi-Supervised Relational Contrastive Learning (SRCL), a model that uses self-supervised contrastive loss and sample relation consistency to better exploit unlabeled data, and demonstrates its effectiveness on the ISIC 2018 skin lesion classification dataset with varying labeled data amounts.
Disease diagnosis from medical images via supervised learning is usually dependent on tedious, error-prone, and costly image labeling by medical experts. Alternatively, semi-supervised learning and self-supervised learning offer effectiveness through the acquisition of valuable insights from readily available unlabeled images. We present Semi-Supervised Relational Contrastive Learning (SRCL), a novel semi-supervised learning model that leverages self-supervised contrastive loss and sample relation consistency for the more meaningful and effective exploitation of unlabeled data. Our experimentation with the SRCL model explores both pre-train/fine-tune and joint learning of the pretext (contrastive learning) and downstream (diagnostic classification) tasks. We validate against the ISIC 2018 Challenge benchmark skin lesion classification dataset and demonstrate the effectiveness of our semi-supervised method on varying amounts of labeled data.