Fig. 7
From: Automatic segmentation of large-scale CT image datasets for detailed body composition analysis
a Illustration of UNET++ model predictions and comparison to reference segmentations for randomly selected CT image examples, from top to bottom (Spleen to SAT) images: from left to right, CT image, ground truth, models predicted output, mask difference between ground truth and prediction, predicted mask overlayed on the original CT image, highlighted segmented region (contour) with mark boundaries. b Illustration of UNET++ model predictions and comparison to reference segmentations for randomly selected CT image examples, from top to bottom (DSAT to Thigh SAT) images: from left to right, CT image, ground truth, models output prediction, mask difference between ground truth and prediction, predicted mask overlayed on the original CT image, highlighted segmented region (contour) with mark boundaries