A modified level set algorithm based on point distance shape constraint for lesion and organ segmentation.

Area of Science:

• Medical Imaging
• Computer Vision
• Computational Anatomy

Background:

• Medical image segmentation is crucial for diagnosis and treatment planning.
• Challenges include noise, intensity inhomogeneity, and unclear boundaries.
• Existing methods struggle with diverse organ and lesion shapes.

Purpose of the Study:

• To develop a flexible point distance shape constraint for level set segmentation.
• To improve the accuracy of segmenting organs and lesions with varied shapes.
• To address limitations of current segmentation techniques in medical imaging.

Main Methods:

• Proposed a novel point distance shape constraint based on Euclidean distance of user-selected points.
• Integrated the shape constraint into a variational level set framework.
• Employed convex relaxation and a primal-dual hybrid gradient algorithm for efficient solving.

Main Results:

• Achieved high segmentation accuracy across multiple datasets: 95.6% (HCC), 94.3% (kidney), 91.3% (prostate MR), 92.7% (lumbar CT), 94.4% (transrectal ultrasound prostate).
• Demonstrated superior performance compared to four state-of-the-art methods through quantitative and qualitative evaluations.
• Validated on diverse medical images including CT, ultrasound, and MRI.

Conclusions:

• The proposed point distance shape constraint segmentation model accurately segments organs and lesions.
• The method is effective for a wide range of shapes in medical images.
• Offers a robust and accurate solution for challenging medical image segmentation tasks.