Statistical Shape Modelling for the Levator Ani

Abstract

Defective pelvic organ support due to injuries of the levator ani is a common problem in women and its intervention requires a thorough understanding of the structure. Three-dimensional surfaces of the levator ani have proved to be a promising method of studying this. In this paper, we propose to build a statistical shape model (SSM) of the levator ani and describe a segmentation technique based on a limited number of control points with the SSM. The SSM was achieved by the use of harmonic shape embedding with the MDL objective function to optimise parameterisation while segmentation was performed by fitting the model to a user defined set of control points. The value of the technique was demonstrated with data acquired from a group of 11 asymptomatic subjects.