Fig. 8
From: GastroHUN an Endoscopy Dataset of Complete Systematic Screening Protocol for the Stomach

Cohen’s Kappa scores comparing model predictions with gastroenterologist labels across all samples in the test set, with the lower triangle showing agreement among experts ("FG” for Fellow Gastroenterologists, “G” for Gastroenterologists) and the diagonal representing model-expert annotations. The trained model was FG (consensus in Table 7).