Figure 1: Ground truth segmentation.

Each data set was segmented manually by two clinical experts, A and B, through the manual outlining of bone structures in each slice (red field). (a) Original CT image slice. (b) Manual ground truth segmentation of clinical expert A (red field).