Adaptive shape prior constrained level sets for bladder MR image segmentation.

Area of Science:

• Medical Imaging
• Image Analysis
• Computational Anatomy

Background:

• Bladder wall thickness measurement using 3D MRI is crucial for detecting abnormalities.
• Accurate segmentation is challenging due to artifacts, weak boundaries, and complex surrounding tissues.

Purpose of the Study:

• To develop an adaptive shape prior constrained directional level set model for segmenting inner and outer bladder walls.
• To refine segmentation using a coupled directional level set model incorporating region information and minimum thickness.

Main Methods:

• Proposed an adaptive shape prior constrained directional level set model.
• Introduced a coupled directional level set model for segmentation refinement.
• Applied the method to 3D bladder MR images.

Main Results:

• Significantly reduced the influence of artifacts within the bladder lumen.
• Effectively mitigated issues caused by complex surrounding tissues.
• Avoided segmentation leakage at weak boundary regions.
• Achieved superior segmentation results compared to existing methods on 11 patient datasets.

Conclusions:

• The proposed method offers improved accuracy and robustness for 3D bladder wall segmentation in MR images.
• This technique enhances the analysis of bladder wall thickness for abnormality detection.
• The model effectively addresses common segmentation challenges in bladder MRI.