The author introduces methods for cross-domain and cross-dimension transfer learning for image-to-graph transformers, demonstrating superior performance in various domains.
Direkte Bild-zu-Graph-Transformation durch Transferlernen über verschiedene Domänen und Dimensionen.