Gradient vector flow based active shape model for lung field segmentation in chest radiographs.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Image Segmentation

Background:

• Accurate segmentation of lung fields in chest radiographs is essential for the reliable detection of various lung pathologies.
• Existing methods, such as Active Shape Models (ASM), face challenges in achieving precise lung field extraction, potentially impacting the accuracy of computer-aided diagnosis (CAD) systems.
• Lung diseases like lung cancer and tuberculosis (TB) require precise diagnostic tools for early intervention.

Purpose of the Study:

• To develop and evaluate a modified Gradient Vector Flow Active Shape Model (GVF-ASM) for enhanced lung field segmentation in chest radiographs.
• To improve the accuracy and robustness of lung field extraction compared to traditional ASM techniques.
• To contribute to the advancement of CAD systems for lung disease diagnosis.

Main Methods:

• Implementation of a modified Gradient Vector Flow (GVF) algorithm integrated with an Active Shape Model (ASM).
• Application of the proposed GVF-ASM technique for automatic lung field segmentation on a dataset of chest radiographs.
• Quantitative comparison of the GVF-ASM segmentation results against standard ASM methods.

Main Results:

• The proposed GVF-ASM technique demonstrated a significant improvement in lung field segmentation accuracy.
• Experimental results indicated an approximate 3-5% enhancement in performance compared to conventional ASM techniques.
• The modified GVF-ASM method proved effective in extracting lung fields with greater precision.

Conclusions:

• The modified GVF-ASM is a promising technique for accurate lung field segmentation in chest radiographs.
• This improved segmentation can enhance the performance of computer-aided diagnosis (CAD) systems for lung diseases.
• The GVF-ASM approach offers a valuable advancement for medical image analysis in pulmonology.