Validation and Implementation of an Automated Boundary Finding Algorithm for Muscle Anatomy Studies
Lansdown D, Ding Z, Damon B
Measurements of muscle volume are important in many biomechanical, applied physiology, and pathophysiology studies. While anatomical MRI provides a precise and accurate means for making these measurements, the need to hand-define regions of interest in multiple imaging slices is subjective and time-consuming. Here, we validate a boundary finding algorithm that uses a combined model of prior shape and smoothness with computer-generated phantoms. The algorithm significantly reduced user input, and ROI quality indices were outstanding. The algorithm was then implemented in anatomical MR images of the leg.