Related Experiment Videos
Cardiac segmentation by a velocity-aided active contour model.
1Samsung Electronics Co. Ltd., Suwon-city, Gyeonggi-do, South Korea. jinsoo.cho@samsung.com
Summary
Accurate heart disease diagnosis relies on precise cardiac segmentation. A new velocity-aided method using orientation gradient force (OGF) and seed contour tracking (SCT) improves endocardial boundary segmentation in phase contrast MRI, despite some artifact limitations.
Area of Science:
- Medical Imaging
- Cardiovascular Imaging
- Biomedical Engineering
Background:
- Accurate cardiac functional analysis is crucial for diagnosing heart disease, relying heavily on precise myocardial boundary segmentation.
- Current segmentation methods struggle with low image quality, complex heart anatomy, and motion, particularly for the endocardial boundary.
- Existing techniques often lack full automation and struggle with clear delineation between myocardium and adjacent structures.
Purpose of the Study:
- To develop and evaluate a novel velocity-aided cardiac segmentation method to enhance the accuracy of myocardial boundary segmentation, especially the endocardial boundary.
- To improve automatic sequential frame segmentation for cardiac MRI analysis.
- To address limitations in current cardiac segmentation techniques using phase contrast MRI data.
Main Methods:
- A modified active contour model incorporating tensor-based orientation gradient force (OGF) was developed for individual frame segmentation.
- An initial seed contour tracking (SCT) algorithm was integrated for automatic sequential frame segmentation.
- The proposed method was validated using phase contrast MRI data from three healthy human volunteers.
Main Results:
- The OGF method improved accuracy and reproducibility of endocardial boundary segmentation, particularly at the lower left ventricle (LV) level and during end systole.
- The SCT algorithm significantly reduced error propagation in sequential frame segmentation, yielding higher accuracy and reproducibility compared to individual frame segmentation.
- Segmentation improvements were less pronounced at the upper LV level and during end diastole; velocity wrap-around artifacts and blood turbulence degraded performance.
Conclusions:
- The developed velocity-aided cardiac segmentation method shows potential for improving myocardial boundary segmentation accuracy using phase contrast MRI.
- The combination of OGF and SCT enhances automated cardiac segmentation, reducing errors in dynamic analysis.
- Future research should focus on mitigating velocity wrap-around artifacts to further optimize cardiac segmentation performance.