Skip to main content
Fig. 7 | European Journal of Hybrid Imaging

Fig. 7

From: Applications of artificial intelligence and deep learning in molecular imaging and radiotherapy

Fig. 7

Examples showing how the automatic Liver-Net results (cyan) were corrected. The contours denote corrected liver masks by radiology residents with 3-year experience (Res3) (blue) and radiology assistant (RA) (yellow). a Case with minor or no corrections: 1.42% for the Liver-Net liver masks, 1.51% (4.3%), and 1.42% (0%) mean relative error (RVE) percentage of corrected slices (CS) for the corrections by Res3 and RA, respectively. b Case where less than half of all slices were corrected: 5.66%, 5.57% (47.6%), and 5.40% (35.7%). c Case where most of the slices were corrected by all observers: 11.48%, 1.62% (74.0%), and 0.10% (78.3%). Reprinted with permission from the Public Library of Science (Chlebus et al. 2019)

Back to article page