Phase symmetry approach applied to children heart chambers segmentation: a comparative study
IEEE Transactions on Bio-Medical Engineering
|April 27, 2011
Summary
This study introduces a fully automatic algorithm for segmenting all heart chambers in echocardiographic images, overcoming speckle noise and artifacts. The novel method significantly improves accuracy compared to existing techniques.
Area of Science:
- Medical Imaging
- Image Processing
- Cardiovascular Ultrasound
Background:
- Echocardiographic image segmentation is challenging due to speckle noise and artifacts.
- Current ultrasound segmentation methods are often semi-automatic, requiring manual contour initialization.
Purpose of the Study:
- To develop a fully automatic algorithm for simultaneous segmentation of all heart cavities.
- To improve upon existing segmentation techniques for echocardiographic images.
Main Methods:
- Algorithm based on phase symmetry approach and level set evolution.
- Utilizes a novel logarithmic-based stopping function for level set evolution.
- Evaluated on echocardiographic images of children.
Main Results:
- The proposed method achieved at least 4% superior performance compared to other methods for segmenting four chambers.
- Demonstrated better performance than watershed and manual segmentation.
- Outperformed other techniques even in challenging cases like right ventricle and left atrium boundary extraction.
Conclusions:
- The developed algorithm offers a robust and accurate solution for fully automatic echocardiographic image segmentation.
- This method provides a significant advancement over existing techniques, particularly for pediatric cardiac imaging.
