Table 3 Features.

From: Distilling knowledge from graph neural networks trained on cell graphs to non-neural student models

Feature names

Features (feature number)

Graph Features

Eccentricity (1), Closeness_of_node (2), Average_Clustering (3), Node_Clustering (4), Sørensen (5), Salton (6), Hub_Promoted (7), Hub_Depressed (8), Centrality (9), Mean_all_neighbors (10), Skew_all_neighbors (11), Kurtosis_all_neighbors (12)

Morphological Features

X (13),Y (14), Contrast (15), Energy (16), Correlation (17), Homogeneity (18), ASM_value (19), Dissimilarity (20), Variance (21), Mean_Image (22), Standard_Deviation (23), Area (24), Major_Axis (25), Minor_Axis (26), Eccentricity_object (27), Perimeter (28), Diameter_object (29), Circularity (30), Mean_convex_hull (31), SD_convex_hull (32)

  1. Note: Skew_all_neighbors and Kurtosis_all_neighbors are computed based on the distribution of edge lengths between neighboring nodes.