Deep learning-based automatic pipeline for quantitative assessment of thigh muscle morphology and fatty infiltration.

The proposed method exhibits excellent accuracy and reproducibility in CSA and FF quantification compared with manual segmentation and can be used in large-scale patient studies.