Skip to main content
. 2022 Aug 23;4(1):vdac138. doi: 10.1093/noajnl/vdac138

Figure 1.

Figure 1.

Segmentation (left column) and detection (right column) agreement between the ground-truth segmentation mask generated by the radiologist and the automatically generated segmentation masks for contrast-enhancing (CE) tumor on a per lesion level (upper row) and a per case level (lower row) within each dataset using violin charts and superimposed box plots. The colors represent each data set.