The segmentation of MR bladder wall in 3D based on minimum closed set model

摘要

Efficient and accurate segmentation of bladder wall on MR images is the most challenging part in constructing the virtual cystoscopy. Segmentation results directly affect the subsequent processing. We present an approach of segmenting the bladder wall by solving minimum-cut problem on closure graphs which was firstly used to solve open-pit mining problem. The results were obtained by optimizing the cost functions designed for individual surface and geometric constraints defining the surface smoothness and interaction. This approach yielded precise results in segmenting the inner border of the bladder wall in 2D space, and has the potential to be extended to higher-dimensional space.