Fig. 3

Architecture of proposed TILs classifier—OralTILs-ViT, which comprises of two feature extractors (\(\phi _{img}\) and \(\phi _{cell}\)), a Fusion module and a Classification head.
Architecture of proposed TILs classifier—OralTILs-ViT, which comprises of two feature extractors (\(\phi _{img}\) and \(\phi _{cell}\)), a Fusion module and a Classification head.