Skip to main content
Insights into Imaging
Fig. 4 | Insights into Imaging

Fig. 4

From: Deep learning-based thigh muscle segmentation for reproducible fat fraction quantification using fat–water decomposition MRI

Fig. 4

MRI images and the segmentation results of an example from Dataset 2 in the testing set. a–c the original water image, fat image and the calculated fat fraction map, d the manual segmentation masks of four ROIs, e, f the preprocessed water image and fat image covering the whole thigh muscles as CNN input channels, g the automated CNN-derived segmentation masks. For d, g, segmentations are superimposed on the water image, with ROI1 (quadriceps femoris) in red, ROI2 (sartorius) in green, ROI3 (gracilis) in blue and ROI4 (hamstrings) in yellow

Back to article page