Fig. 2: Map comparison.
From: Deep learning quantifies pathologists’ visual patterns for whole slide image diagnosis
Between the pathologists’ pixel-level manual annotations for identifying tumor regions (a), heatmaps representing the pathologists’ selected ROIs (b), heatmaps generated by PEAN as the pathologists’ “expert knowledge” (c), and heatmaps depicting the attention trajectories generated by PEAN imitating the pathologists’ slide-review behavior (d).
