Skip to main content
. 2022 Mar 14;9(2):027501. doi: 10.1117/1.JMI.9.2.027501

Fig. 7.

Fig. 7

Comparison of confusion matrices of (a) superpixel-based approach and (b) patch-based approach. The rows represent the ground truth class-labels and the columns represent the predictions. Due to high imbalances in the number of pixel per class, a relative representation of the confusion matrix was chosen.