Novel approach to segment the inner and outer boundaries of the bladder wall in T2-weighted magnetic resonance images.

Diagnosis of bladder-related conditions needs critical measurements which require the segmentation of the inner and outer boundaries of the bladder wall. In T2-weighted MR images, the low-signal intensity bladder wall can be identified due to the large contrast with the high-signal intensity urine and perivesical fat. In this article, two deformable models are proposed to segment the bladder wall. Based on the imaging features of the bladder, a modified geodesic active contour is proposed to segment the inner boundary. This method uses the statistical information of the bladder lumen and can handle the intensity variation in MR images. Having obtained the inner boundary, a shape influence field is formed and integrated with the Chan-Vese (C-V) model to segment the outer boundary. The shape-guided C-V model can prevent the overlapping between the two boundaries when the appearance of the bladder wall is blurred. Segmentation examples are presented and analyzed to demonstrate the effectiveness of this novel approach.