Segmenting echocardiography images using B-Spline snake and active ellipse model.

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Image Segmentation

Background:

• Echocardiography is crucial for assessing cardiac function.
• Accurate segmentation of the Left Ventricle (LV) is essential for quantitative analysis.
• Existing segmentation methods face challenges with myocardial boundary gaps.

Purpose of the Study:

• To develop a fully automated method for LV segmentation in echocardiography.
• To introduce an active ellipse model for precise LV chamber delineation.
• To improve segmentation accuracy and efficiency compared to prior techniques.

Main Methods:

• Development of a novel active ellipse model for automatic LV chamber ellipse detection.
• Utilizing a modified B-Spline Snake algorithm initialized with the active ellipse.
• Addressing and overcoming issues of myocardial boundary gaps in segmentation.

Main Results:

• The active ellipse model successfully identifies the optimal ellipse within the LV.
• The B-Spline Snake algorithm, initialized by the ellipse, provides robust segmentation.
• Achieved a Dice's coefficient of 92.30 ± 4.45% on 20 patient datasets.
• Demonstrated superior speed and accuracy compared to existing methods.

Conclusions:

• The proposed automated method offers a significant advancement in echocardiography image segmentation.
• The active ellipse model effectively resolves challenges posed by myocardial boundary gaps.
• This technique provides a reliable and high-performing solution for LV segmentation.