Fig. 1From: Deep semi-supervised learning for brain tumor classificationThe proposed deep semi-supervised learning scheme for glioma classification, where \(\mathcal {L}\), \(\mathcal {U}\) and \(\mathcal {T}\) denote the labeled training dataset, unlabeled training dataset and the testing dataset, \(\mathcal {Z}\) denotes the feature set, and \(\left \{\hat {y}_{j}\right \}\) represents the estimated labels for images in \(\mathcal {U}\)Back to article page