Skip to main content
Fig. 1 | BMC Medical Imaging

Fig. 1

From: Enhancing basal cell carcinoma classification in preoperative biopsies via transfer learning with weakly supervised graph transformers

Fig. 1

Flow chart of the key ideas in the model. The WSI is first divided into patches, and for each patch, a feature vector is generated using SimCLR. Graphs are then constructed to incorporate spatial information into each feature vector. These graphs are passed through the graph-transformer, which utilizes max pooling to reduce dimensionality before processing in the vision transformer

Back to article page

BMC Medical Imaging

ISSN: 1471-2342

Contact us