Automatic Liver Segmentation from CT Images Using Single-Block Linear Detection.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Radiology

Background:

• Automatic liver segmentation is crucial for liver disease analysis, yet challenging due to anatomical variations.
• Existing segmentation methods require improvement in precision and efficiency.
• Manual segmentation is time-consuming and resource-intensive.

Purpose of the Study:

• To propose a fast and robust algorithm for automatic liver extraction from CT images.
• To enhance segmentation precision and reduce computational complexity.
• To evaluate the algorithm's performance on normal and abnormal liver CT scans.

Main Methods:

• A novel, non-iterative algorithm utilizing single-block linear detection for liver extraction.
• The method does not rely on critical initialization, enhancing robustness.
• Implementation focused on reducing computational time and complexity.

Main Results:

• Achieved high performance in segmenting normal and abnormal liver CT images (hemangioma, cancer).
• For liver cancer, average sensitivity, accuracy, and specificity were 96.59%, 98.65%, and 99.03%, respectively.
• Segmentation results closely approximated manual segmentation by expert radiologists.

Conclusions:

• The proposed algorithm offers a fast, robust, and precise solution for automatic liver segmentation.
• It demonstrates superior flexibility and comparable performance to recent methods.
• This technique has the potential to reduce costs and improve the efficiency of liver disease analysis.