Table 3 F1 scores of OMG-Net and the top 3 models from the MIDOG 2022 challenge in the overlap subsets between the MIDOG++ and MIDOG 2022 test sets

From: A deep learning framework deploying segment anything to detect pan-cancer mitotic figures from haematoxylin and eosin-stained slides

Tumour types

OMG-Net

TIA centre34

TCS research35

USZ/UZH Zurich36

Melanoma

0.85

0.80 [0.74,0.84]

0.76 [0.66,0.80]

0.79 [0.74,0.83]

Cutaneous mast cell tumoura

0.87

0.83 [0.81,0.86]

0.76 [0.58,0.83]

0.73 [0.66,0.79]

  1. aCanine Specimens.