Homogeneity- and density distance-driven active contours for medical image segmentation

Area of Science:

• Medical image analysis
• Computer vision
• Biomedical engineering

Background:

• Medical image segmentation is crucial for diagnosis and treatment planning.
• Existing active contour (AC) models face challenges with inhomogeneous objects and low-contrast images.
• Objects like bones exhibit high internal intensity variation, complicating segmentation.

Purpose of the Study:

• To develop a novel active contour (AC) model for enhanced medical image segmentation.
• To address limitations of conventional methods in segmenting inhomogeneous and low-contrast medical images.
• To improve the accuracy of segmenting distinct organs and structures in medical imaging datasets.

Main Methods:

• A new active contour (AC) model was developed using a convex combination of two energy functionals.
• The model aims to simultaneously minimize internal object inhomogeneity and maximize object-background distance.
• Performance was evaluated on computed tomography (CT) and magnetic resonance imaging (MRI) datasets.

Main Results:

• The proposed AC model demonstrated comparable performance to the conventional Chan-Vese model on CT images.
• The novel model showed superior performance on MRI datasets, which typically have lower contrast.
• The combined energy functional approach effectively handled object inhomogeneity and background separation.

Conclusions:

• The novel active contour model offers improved performance for medical image segmentation, particularly in challenging low-contrast scenarios like MRI.
• The convex combination of energy functionals provides a robust approach for segmenting inhomogeneous medical image objects.
• This method enhances the accuracy of distinguishing between different anatomical structures in medical imaging.