Fig. 7From: Pelvic U-Net: multi-label semantic segmentation of pelvic organs at risk for radiation therapy anal cancer patients using a deeply supervised shuffle attention convolutional neural networkTime-based evaluation result comparing the mean manual delineation time for all test patients (\({\text{n}} = 15\)) against the time needed to manually correct model predictions and model prediction time (without corrections). Mean manual correction times per structure are shown in the Box and Whisker plot (a). Mean segmentation time could be reduced from 40Â min to about 12Â min per patient using manually corrected model predictions, and to 4Â min, using predictions of the Pelvic U-Net only (b)Back to article page