Development of 3D models of pelvic floor muscles based on deformable models

Abstract

Pelvic floor dysfunction is a hidden problem that is widely unknown. Statistics show that 1 in every 10 women will suffer from pelvic floor dysfunction so severely that they may need surgery. Furthermore, statistics show that 30 to 40% of women experience varying degree of urinary incontinence in their lifetime. Other study has found that the prevalence of this problem among nulliparous women ranges from 8 to 32% and increases with age.Magnetic resonance imaging (MRI) has been used in the diagnostic evaluation of the pelvic floor dysfunctions. Static images can display the morphology of the pelvic floor while dynamic images can show their functional changes that occur on straining and contraction. In the past, MR images are segmented manually in order to generate the 3D solid models of the pelvic floor muscles. However, manual segmentation is very time-consuming. Besides, manual results are not reproducible and suffer from intra-observer and inter-observer variability.In the recent few decades many effective algorithms have been proposed to perform the computer-aided segmentation. Of these, deformable models have been achieving great success in the segmentation of complex structures from medical images.In the presented work algorithms based on deformable models are used to segment the female pelvic floor.