An Improved Fuzzy Connectedness Method for Automatic Three-Dimensional Liver Vessel Segmentation in CT Images

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Image Processing

Background:

• Accurate segmentation of liver vessels in computed tomography (CT) images is crucial for diagnosis and surgical planning.
• Traditional fuzzy connectedness (FC) methods often rely on intensity information, which can be insufficient for complex vessel structures.
• Existing segmentation techniques may require multiple manual seeds, increasing procedural complexity and time.

Purpose of the Study:

• To develop an improved fuzzy connectedness (FC) method for automatic 3D liver vessel segmentation in CT images.
• To enhance the fuzzy affinity function by incorporating vesselness information.
• To automate the initialization process, reducing the need for multiple user-defined seeds.

Main Methods:

• An improved fuzzy connectedness (FC) algorithm was developed, integrating a vessel-enhanced (vesselness) image into its fuzzy affinity function.
• A novel vesselness filter was designed, featuring adaptive sigmoid filtering and background suppression.
• Automatic initialization of the FC fuzzy scene was achieved using the Otsu segmentation algorithm and a single adaptive seed.

Main Results:

• The improved FC method demonstrated superior performance compared to traditional FC, region growing, and threshold level set methods.
• On the 3Dircadb dataset, the method achieved an average accuracy of (96.4 ± 1.1)%, sensitivity of (73.7 ± 7.6)%, specificity of (97.4 ± 1.3)%, and Dice coefficient of (67.3 ± 5.7)%.
• On the Sliver07 dataset, performance metrics were: accuracy (96.8 ± 0.6)%, sensitivity (89.1 ± 6.8)%, specificity (97.6 ± 1.1)%, and Dice coefficient (71.4 ± 7.6)%.

Conclusions:

• The proposed improved fuzzy connectedness method offers a robust and automated solution for 3D liver vessel segmentation from CT images.
• The integration of vesselness information and automated initialization significantly enhances segmentation accuracy and efficiency.
• This improved FC method holds potential as a valuable tool for clinical applications in liver imaging analysis.