Skip to main content
. 2024 Jun 15;485(6):1095–1105. doi: 10.1007/s00428-024-03841-5

Fig. 2.

Fig. 2

AI(H) predictions on H&E slides. The analysis of images is a multi-step process. A H&E-stained WSIs are used for detections (scale bar is 500 µm). B The AI model begins with segmenting liver tissue into its normal micro-structures, such as portal area, parenchyma, and central vein (scale bar is 500 µm). C The relevant necroinflammation features are detected in hepatitis landscape (scale bar is 250 µm). D Total five classes of immune cells (lymphocytes, plasma cells, macrophages, neutrophils, eosinophils) are detected and classified in all over the liver tissue. (scale bar is 200 µm). AI(H), artificial intelligence for hepatitis; H&E, hematoxylin and eosin; WSI, whole-slide image