Skip to main content
. 2022 Jan 24;11:782515. doi: 10.3389/fonc.2021.782515

Figure 2.

Figure 2

Deep learning model. Deep learning model. (A)The whole-slide images were tiled into 256×256-pixel tiles and dropped the tiles with low amount of information while tiling. (B) Tiles before and after color normalization. (C) The convolutional neural network was trained as a tumor detector at tile level. (D) The model training for EMT subtype classification at tile level. (E) EMT subtype prediction at patient level. (F) Distribution of the number of tiles generated per case.