Fig. 7
From: Artificial intelligence in neuropathology: deep learning-based assessment of tauopathy
FCN performance in WSI compared to expert annotated ground truth. High (a–d) and low (e–h) background sections showed similar results. a, e Naive digitized slide, AT8 IHC. b, f Expert annotated ground truth. c, g FCN detection of NFTs. d, h Overlapping image of the expert call annotations and FCN performance. The concordance of calls is used to assess true positives (TP) calls, false positives (FP), and false negatives (FN) for further calculation of FCN performance metrics
